We are working on some future high speed routers with around 100Gbit throughput, and we would like your input on what types of ports you would want to see in such devices.
If you imagine a device with around 16 ports (just an example), how many of them would you want to be (just examples, you can add your own ideas):

SFP+ (10G)
SFP28 (25G)
QSFP (4x10G)
10Gbit copper (10G)
Management with PoE

Let us know in the replies here

Only some ideas:

Router A: 16x SFP+
Router B: 16x SFP+/16x RJ45 10GbE (combo ports)
Router C: 8x SFP+/8x RJ45 10GbE (combo ports) + 2x QSFP
:
:
later
(Router X: 24x SFP+)
(Router Y: 24x SFP+ plus 4x QSFP)
(Router Z: 24x SFP+ plus 4x QSFP28)

(Speeking for myself: I do not need 10GbE copper at all as 10GbE SFP+ optics are cheap nowadays)

all with:
- dual power supplies, hot swappable
- 1x RJ45/SFP combo management interface

I didn't know what ASIC you are looking at, but I'd prefer an ASIC/SOC with less CPUs (1big one instead of 72 small ones) According to that slide from Cavium, the biggest highend ASIC is only 500$, so it would be absolutely possible to build an inexpensive highend router, although this means multiple CPUs and therefore single thread performance isn't that great.

The Indigp NPS-400 Network Processor by EZchip looks better. It looks like it has less CPUs. I cannot find pricing information, unfortunately

But you talked about 100Gbit/s throughput, so the NP-5 seems to be the right one:

I think that would be a great idea some group of port be on a expansion slot so we can change the type depending on our needs.



Enviado de meu XT1580 usando Tapatalk

I think that would be a great idea some group of port be on a expansion slot so we can change the type depending on our needs.

What format do you mean?

I think that would be a great idea some group of port be on a expansion slot so we can change the type depending on our needs.

I think that would raise the costs for those routers too much. Fixed configurations would be a less expensive solution. 1x product design / mainboard with different ports attached to the ASIC.

I think that would be a great idea some group of port be on a expansion slot so we can change the type depending on our needs.

I think that would raise the costs for those routers too much. Fixed configurations would be a less expensive solution. 1x product design / mainboard with different ports attached to the ASIC.

Router with 24 port? Really it not a over port config for a router?

I think a version with expansion slot is not bad, it will be a super core router, I don't think that will make super high cost on the product



Enviado de meu XT1580 usando Tapatalk

I think that would be a great idea some group of port be on a expansion slot so we can change the type depending on our needs.

I think that would raise the costs for those routers too much. Fixed configurations would be a less expensive solution. 1x product design / mainboard with different ports attached to the ASIC.

Router with 24 port? Really it not a over port config for a router?

I think a version with expansion slot is not bad, it will be a super core router, I don't think that will make super high cost on the product



Enviado de meu XT1580 usando Tapatalk

You can already use a high end PC as your router.

I think that would be a great idea some group of port be on a expansion slot so we can change the type depending on our needs.

I think that would raise the costs for those routers too much. Fixed configurations would be a less expensive solution. 1x product design / mainboard with different ports attached to the ASIC.

Router with 24 port? Really it not a over port config for a router?

I think a version with expansion slot is not bad, it will be a super core router, I don't think that will make super high cost on the product



Enviado de meu XT1580 usando Tapatalk

You can already use a high end PC as your router.

But do you agree that is not they same of a dedicated hardware for this
I really don't think that someone will build a high end server to be placed on the the core high density network traffic


Enviado de meu XT1580 usando Tapatalk

The advantage of specifically manufactured hardware is that manufacture can test it at the specific configuration. If you take that away, is becomes the same as any other PC where you have no driver support.

Primo.
Is it possible to just make funcionality of stacking Mikrotik's devices ? High-end ones. Then we can buy models with proper ports and stack them. No need for zyllions versions ....

Secundo.
One mainboard and slots for different daughter cards .... are we going to be CISCO/Juniper like ?

Primo.
Is it possible to just make funcionality of stacking Mikrotik's devices ? High-end ones. Then we can buy models with proper ports and stack them. No need for zyllions versions ....

Secundo.
One mainboard and slots for different daughter cards .... are we going to be CISCO/Juniper like ?

And this is a problem?


Enviado de meu XT1580 usando Tapatalk

What's wrong with giving the big boys a run for their money?
In an other post, I had suggested a chassis with slots, like the Cat6500 family.
In previous MUMs, consultants have talked about replacing Cisco with MikroTik boxes.

I think MikroTik is going on the right direction.

Stackable is also a great idea. Grow your network with your business.

Off topic, MSTP (pvstp, 802.1Q) pleeease! If you're going big, go all the way.

Sent from Tapatalk

...One mainboard and slots for different daughter cards .... are we going to be CISCO/Juniper like ?

And this is a problem?

Not at all.... havn't you spotted ?

Stacking could be even the first step to HA funcionality which is missing for ROS.

I dont think high end routers should be based on Cavium processors.
They have low freq upto 1.2Ghz and some specific RouterOS tasks require higher clock speeds such as simple queues, BGP, PPP etc..

So unless RouterOS is highly multi-threaded in such scenarios like to be used as a PPP concentrator etc.., I would prefer some different architecture than Cavium which will provide higher clock speeds instead of super-high multiple cores with low clock speeds.

...One mainboard and slots for different daughter cards .... are we going to be CISCO/Juniper like ?

And this is a problem?

Not at all.... havn't you spotted ?

Stacking could be even the first step to HA funcionality which is missing for ROS.

Ah ok
I thought that you was complaining....

Ha and some kind of Hsrp for mk

Enviado de meu XT1580 usando Tapatalk

New inexpensive member to the CCR family of routers:

- five 1Gb ports
- five 10Gb ports
- four SFP+ ports
- two USB 3.0 ports

No port sharing / TILE-Gx based CPU

12 x SFP+ (10G)
4 x QSFP (4x10G)
1 x Management (no PoE)
1 x serial management

48 port Gb/e and 4x 10Gb/e (SFP+) (TileGX based)
24 port Gb/e and 2x 10Gb/e (SFP+) (TileGX based)
16/32 port 10Gb/e (SFP+) (TileGX based)

The above *NEED* to be able to do line rate for IP/IPv6/MPLS traffic, or close enough to it (like the way UBNT does by using an Octeon ASIC to get a decent level of mpps, although UBNT doesn't do MPLS at the mpps level. We need MPLS at that level too).

Or please support the following chipsets from Broadcom. The Trident II and the Tomahawk.

https://people.ucsc.edu/~warner/Bufs/St ... tation.pdf <-- Trident II
https://www.broadcom.com/products/ether ... c/bcm56850 <-- Trident II
https://www.broadcom.com/products/ether ... c/bcm56960 <-- Tomahawk

Lastly, if one were to compare to say Juniper or Cisco I would say the comparisons would be...

A 'tik equivalent to the Juniper ACX1100, or an equivalent to the Cisco ASR-920-12SZ-IM.

A 'tik equivalent to the Juniper ACX5048 would also be fantastic.

Personally, I'd be interested 2x QSFP and the rest with SFP+ and the obligatory Management Port paired with a rs232 for console access.

The pair of QSFP would provide redundant access to e.g. core switches, while the SFP+ would connect to ISP, DMZ, or further routers/firewalls/gateways.

In regards to the PoE-in functionality, I very much doubt that even PoE+ will provide sufficient power. PoE++ / 4PPoE on the other hand just might. I'd much rather see a pair of redundant hotswappable PSUs with a bit of headspace.
Speaking of hotswappable: I think we'd all love hotswappable fans.

Grüße,
Philip

2x QFP28 (with breakout support), 10x SFP+
2x QFP28 (with breakout support), 3x QFP (with breakout support)
3x QFP (with breakout support)

QFP28 Can do 1x100G or can be broken out into multiple 25G.
QFP can do 1x 40G or can be broken out into multiple 10G.

+1 for SFP28 and QSFP28, maybe a 4 port router

Great thread!

25 and 40 gig would be my focus. I'll post a longer response when I have a spare minute to think about specific port configurations. Right now, I've got to get ready for the US MUM

48 port Gb/e and 4x 10Gb/e (SFP+) (TileGX based)
:
Or please support the following chipsets from Broadcom. The Trident II and the Tomahawk.
.

- Trident/Tomahawk is meant for switches not for router.

- Does TileGX has a future at Ezchip/Mellanox? There are other options: http://www.mellanox.com/page/npu_overview (NP/NPS):

The ccr1036 with two SFP+ (10G) ports based on speed tests is working great.
viewtopic.php?f=2&t=104266

Going forward more ports is better and support for QSFP ports would allow additional networking options. I have not tested the CHR yet beyond the SFP+ (10G) ports in my btest CHR but suspect scaling is allowed based on licensing, currently a p10.
SFP+ (10G)
QSFP (4x10G)

A side question to Normis and team: any opportunity for routerboard designs with native x86 or x64 support. Tile has been good as a platform but i do also like the x64 potential in the CHR guests. And one more shout out for CHR vmware-tools

Re: Which types of ports would you like to see for a high speed router

I really like this topic !!!

My suggestions are:

- Possible special stack ports
Where there is a minimum of two stack ports per device (or four ports) , which enables the ability to stack more than two Mikrotik devices.
Where there is a new Mikrotik stack interface using normal copper RJ-45 connectors and cat6 compatible (and possible cat7 or cat8 compatible.
Where stacking could enable two stackable Mikrotik routers and possibly two stackable Mikrotik switches to act and behave as a single device.
Where there could be an option that if these copper RJ-45 stack ports are not used in a stack configuration, then they can alternately be used in a 10-gig copper Ethernet environment.
With high-speed stack ports, it would then be possible for Mikrotik to offer a lower-cost basic system, then additional Mikrotik devices could be ordered later, or as/when new Mikrotik stackable devices come out, then can also be added to the stack.
Where with the base system, it is possible to bring up a small--to-medium--to-large number of interfaces by adding to the Mikrotik stack interfaces.
Where a 48 port 10/100/1000 Ethernet option is simply a new stacked device.
Where a 12 to 48 port SFP interface is simply a new stacked device.
Where a 12 to 48 port SFP+ interface is simply a new stacked device.
Where additional 12 port 10-gig copper Ethernet interfaces can be added into the stack.


- The ability to perform true bonded/load-sharing of SFP+ or 10-gig-copper ports to achieve 20 t 40 Gig throughput (True single MAC address to single MAC address balanced load-sharing)
Where it would be possible to have up to four (or more) 10-gig fiber (or copper) which are remotely separated to have greater than 10-gig throughput.

- SFP+ ports are backwards compatible with normal SFP devices ( This would be assumed )

- Possibly four 1-gig Ethernet ports (with POE on each port)

- For faster than 10-gig interfaces (not including stack ports), I would like see the most common/compatible/popular/standard interface used.
I have no experience with SFP28 or QSFP)

- CPU
A bad-ass and popular CPU (Intel XEON), with lots of built-in CPU internal cache
Then as newer & faster CPUs come out, it is just a simple matter of switching to the newer CPU
High-speed RAM (standard high-speed normal memory sticks) (not motherboard surface mounted - user upgradeable & user replaceable)

- Possibly one to four PCI card slots for future cards/interfaces.
There may be a desire to use an already existing PCI card in a Mikrotik - where a quick/simple ROS driver could bring up the PCI card device/interface.

- Dual power supplies (the new system should be able to be powered by industry standardTelco-48-Volts, and/or standard AC outlets
Where the AC power supplies are NOT built-into the new Mikrotik - but instead the AC power supplies are standard telco-48-volts out and then connected to the new Mikrotik.
Were 12 or 24 Volts is not needed or used (except to POE ports)

- One serial port

- All flash ram on removable user replaceable industry standard media

- ROS (Intel/AMD) 64-bit firmware

- Future software/firmware ability for two independent systems can function in a hot-stand-by-redundant mode (something like Cisco HSRP)

- Future ROS firmware/software for true VRF interfaces/routing tables (similar to how Cisco handles VRF interfaces/routes without injecting routed/IP-addresses/MAC-addresses into the entire system.
Where it would be additionally possible to also have a VRF interface for true out-of-band management interface on a router.
Where multiple independent VRF routing tables and interfaces to not see each other and are not seen by the base system.
With VRF features, I suspect a MetaRouter feature will not be needed or even desired.
These new VRF features should support the normal stuff a network configuration will probably encounter - such as VRF to VRF bridges & VRF DHCP & VRF NAT & ability to bridge VRF interfaces to other VRF configurations & bridge VRF interfaces to live interfaces & bridge VRF interfaces to non-VRF interfaces.
Possible idea - Ability CPU limit VRF functions so that a single or multiple or all combined VRF features do not degrade the primary non-VRF configuration (UNIX nice )

- This is a MUST ... The ability to perform full BGP and quickly handle all BGP routes (without the need to upgrade something at the BGP tables increase and more BGP interfaces/peers are added.
Possible idea ... Ability to run full BGP under a VRF environment instead of under the non-VRF environment --- there can be some huge advantages to a VRF BGP environment. And possibly the ability to run multiple independent full BGP environments to multiple up-stream connectivity peering networks.

- No excessive fancy LEDS. Possibly just a small simple LED display which can actually show real-time diagnostic information (verses individual single LET lights)

- Two platforms at first.
A low cost (fewer interfaces & devices mounted to the motherboard. Something most of us -and- our customers can afford without breaking our budgets.

A high-end high-throughput version (exact same motherboard - just with faster parts and all the interfaces mounted on the motherboard.
This simplifies production and the low-cost version gets the product out in the field quicker to the engineers who want to test-drive it in a LAB or evaluate in production.

- Standard rack mount chassis

- No huge big blade chassis in the design - where expansion and additional devices are added to the system using Mikrotik stack interfaces (keep costs down to get started)

- When this low-cost stackable configuration starts to take off and becomes desired and popular, then a blade version of the same thing could then be latter designed which would then use a high-speed back-plane between cards.
By starting with a stack interface (instead of a blade back-plane), the initial costs would be lower.
Later, when a pattern and market develops for an even faster version, then the basic designs could then be slightly re-engineered to use a back-plane in a large chassis (which is more expensive/faster).
Stacking existing and older Mikrotik devices (some with less memory and flash and slower CPUs) could be possibly done with a RO-Stack OS, where the new OS has everything but the minimum necessary software drivers to enable the new RO-Stack OS to participate in the stack and deliver the interfaces over an Ethernet stack interface. This would add value to all existing Mikrotik devices because a fork-lift replacement would not be necessary to get started with this new "Mikrotik stack head-end device".
Consider a CHR stack version of ROS. It might be interesting to combine other CHR systems to be able to participate in the stack (possible VRF BGP via stacked CHR ??? Interesting idea ? ) I would want to name it CHR-Stack (over Ethernet interfaces)

- With a stack interface (using a normal Ethernet interface), it would be great to have all existing Mikrotik ROS devices support 1-Gig or 10-Gig stacking. This adds value to all existing Mikrotik products already in production and already in use.

EDIT/add - If stack interfaces are considered - the stack interfaces do not need to support power over the stack. Each stack device should use it's own (48 volt or AC) power supply.
EDIT/add - A possible hardware reset-/-reboot watchdog. The ROS (or what I would like to call it the RO-Stack OS) could do something like a Netwatch IP ping or something) to keep resetting/clearing the hardware watchdog. If the ROS ( or RO-Stack OS ) locks up - crashes, then the hardware watchdog will continue counting down to zero which could then trigger a full-blown reboot - resulting in an auto-recovered system. The slave stacked devices could also participate in the primary watchdog. Possibly by using a reset wire in the RJ-45 stack cable. Thus, a reboot command on the primary master stack device could possibly reboot everything in the stack if wanted/needed.
EDIT/add - Clarification - When I refer to stacking using Ethernet ports , I am not referring to stacking using IP. It could be something like MAC to MAC. Where the master stacked devices knows the MAC address of everything that is supposed to be in the stack and what it is. The hardware Ethernet switch chips (or ROS software bridge if necessary on old devices) could be used to re-distribute the stack data at wire-speed to other stack ports. It might be even possible to allocate/use a special Vlan for the Ethernet connector(s) stack interfaces.
EDIT/add - Optional - faster than 10-gig stack ports could also be combined and also work with 10-gig or 1-gig interfaces. Possibly, the administrator simply just defines what ports are used for stacking instead of being normal interfaces. This makes it flexible for how the ports are used. With user/administrator selectable stack ports and with the ability for the master stack device to talk to different stacked slave devices at different speeds, the system would never become obsolete - instead it just becomes bigger faster and more desirable for commercial business use/installations.
EDIT/add - slaves on the stack only need to be configured to know the MAC address (or auto-discover) the master slave interface. The master stack device should hold the entire configuration for everything and be as simple as making a backup on any other Mikrotik.

North Idaho Tom Jones

Re: Which types of ports would you like to see for a high speed router

I really like this topic !!!

My suggestions are:

- Possible special stack ports
Where there is a minimum of two stack ports per device (or four ports) , which enables the ability to stack more than two Mikrotik devices.
Where there is a new Mikrotik stack interface using normal copper RJ-45 connectors and cat6 compatible (and possible cat7 or cat8 compatible.
Where stacking could enable two stackable Mikrotik routers and possibly two stackable Mikrotik switches to act and behave as a single device.
Where there could be an option that if these copper RJ-45 stack ports are not used in a stack configuration, then they can alternately be used in a 10-gig copper Ethernet environment.
With high-speed stack ports, it would then be possible for Mikrotik to offer a lower-cost basic system, then additional Mikrotik devices could be ordered later, or as/when new Mikrotik stackable devices come out, then can also be added to the stack.
Where with the base system, it is possible to bring up a small--to-medium--to-large number of interfaces by adding to the Mikrotik stack interfaces.
Where a 48 port 10/100/1000 Ethernet option is simply a new stacked device.
Where a 12 to 48 port SFP interface is simply a new stacked device.
Where a 12 to 48 port SFP+ interface is simply a new stacked device.
Where additional 12 port 10-gig copper Ethernet interfaces can be added into the stack.


- The ability to perform true bonded/load-sharing of SFP+ or 10-gig-copper ports to achieve 20 t 40 Gig throughput (True single MAC address to single MAC address balanced load-sharing)
Where it would be possible to have up to four (or more) 10-gig fiber (or copper) which are remotely separated to have greater than 10-gig throughput.

- SFP+ ports are backwards compatible with normal SFP devices ( This would be assumed )

- Possibly four 1-gig Ethernet ports (with POE on each port)

- For faster than 10-gig interfaces (not including stack ports), I would like see the most common/compatible/popular/standard interface used.
I have no experience with SFP28 or QSFP)

- CPU
A bad-ass and popular CPU (Intel XEON), with lots of built-in CPU internal cache
Then as newer & faster CPUs come out, it is just a simple matter of switching to the newer CPU
High-speed RAM (standard high-speed normal memory sticks) (not motherboard surface mounted - user upgradeable & user replaceable)

- Possibly one to four PCI card slots for future cards/interfaces.
There may be a desire to use an already existing PCI card in a Mikrotik - where a quick/simple ROS driver could bring up the PCI card device/interface.

- Dual power supplies (the new system should be able to be powered by industry standardTelco-48-Volts, and/or standard AC outlets
Where the AC power supplies are NOT built-into the new Mikrotik - but instead the AC power supplies are standard telco-48-volts out and then connected to the new Mikrotik.
Were 12 or 24 Volts is not needed or used (except to POE ports)

- One serial port

- All flash ram on removable user replaceable industry standard media

- ROS (Intel/AMD) 64-bit firmware

- Future software/firmware ability for two independent systems can function in a hot-stand-by-redundant mode (something like Cisco HSRP)

- Future ROS firmware/software for true VRF interfaces/routing tables (similar to how Cisco handles VRF interfaces/routes without injecting routed/IP-addresses/MAC-addresses into the entire system.
Where it would be additionally possible to also have a VRF interface for true out-of-band management interface on a router.
Where multiple independent VRF routing tables and interfaces to not see each other and are not seen by the base system.
With VRF features, I suspect a MetaRouter feature will not be needed or even desired.
These new VRF features should support the normal stuff a network configuration will probably encounter - such as VRF to VRF bridges & VRF DHCP & VRF NAT & ability to bridge VRF interfaces to other VRF configurations & bridge VRF interfaces to live interfaces & bridge VRF interfaces to non-VRF interfaces.
Possible idea - Ability CPU limit VRF functions so that a single or multiple or all combined VRF features do not degrade the primary non-VRF configuration (UNIX nice )

- This is a MUST ... The ability to perform full BGP and quickly handle all BGP routes (without the need to upgrade something at the BGP tables increase and more BGP interfaces/peers are added.
Possible idea ... Ability to run full BGP under a VRF environment instead of under the non-VRF environment --- there can be some huge advantages to a VRF BGP environment. And possibly the ability to run multiple independent full BGP environments to multiple up-stream connectivity peering networks.

- No excessive fancy LEDS. Possibly just a small simple LED display which can actually show real-time diagnostic information (verses individual single LET lights)

- Two platforms at first.
A low cost (fewer interfaces & devices mounted to the motherboard. Something most of us -and- our customers can afford without breaking our budgets.

A high-end high-throughput version (exact same motherboard - just with faster parts and all the interfaces mounted on the motherboard.
This simplifies production and the low-cost version gets the product out in the field quicker to the engineers who want to test-drive it in a LAB or evaluate in production.

- Standard rack mount chassis

- No huge big blade chassis in the design - where expansion and additional devices are added to the system using Mikrotik stack interfaces (keep costs down to get started)

- When this low-cost stackable configuration starts to take off and becomes desired and popular, then a blade version of the same thing could then be latter designed which would then use a high-speed back-plane between cards.
By starting with a stack interface (instead of a blade back-plane), the initial costs would be lower.
Later, when a pattern and market develops for an even faster version, then the basic designs could then be slightly re-engineered to use a back-plane in a large chassis (which is more expensive/faster).
Stacking existing and older Mikrotik devices (some with less memory and flash and slower CPUs) could be possibly done with a RO-Stack OS, where the new OS has everything but the minimum necessary software drivers to enable the new RO-Stack OS to participate in the stack and deliver the interfaces over an Ethernet stack interface. This would add value to all existing Mikrotik devices because a fork-lift replacement would not be necessary to get started with this new "Mikrotik stack head-end device".
Consider a CHR stack version of ROS. It might be interesting to combine other CHR systems to be able to participate in the stack (possible VRF BGP via stacked CHR ??? Interesting idea ? ) I would want to name it CHR-Stack (over Ethernet interfaces)

- With a stack interface (using a normal Ethernet interface), it would be great to have all existing Mikrotik ROS devices support 1-Gig or 10-Gig stacking. This adds value to all existing Mikrotik products already in production and already in use.

North Idaho Tom Jones

OMG this is big text
But Iiked [emoji6]

Enviado de meu XT1580 usando Tapatalk

IMHO, there is no point in developing anything new with 10Gbps/40Gbps ports at this point in time and the new 16 x SFP+ switch you presented in Milan this year should have already been made with SFP28 ports. But let's keep it simple for starters...

• Considering that new SFP28 ports should be able to support old SFP+ modules, how about a CCR2572 with 8 x SFP28 ports as a direct in-place upgrade for CCR1072?
• Then bring it up a notch and create something like CCR10072 (yes, that's an extra zero) with 8 x SFP28 + 1 (or 2?) x QSFP28, although it would probably need a better CPU.
• Then let's talk datacenters - how about a new 16 x SFP28 + 2 x QSFP28 switch with enough CPU power for some OSPF/iBGP routing experiments?

Naturally, this would also require some new SFP28 modules, SFP28/QSFP28 DAC cables and so on...

Edit: It looks like some people here aren't aware that new 25/100 Gbps ports can also support "old" 10/40 Gbps speeds? Here's some reading material:

• https://www.mellanox.com/blog/2016/03/2 ... w-amazing/
• https://www.cisco.com/c/en/us/products/ ... 36651.html

So IMHO the conclusion so far is:

Hardware
1. Stacking as the fastest path for extension
2. Rack chassis with highspeed backplane to connect interfaces and CPU motherboards tailored for current needs
3. True HA funcionality

If we could have CPU-board with ROS kernel separated from interface cards then we can easily upgrade router without changing any rule.
If the interface name will be local and bound to interface card id then rules based on that name will need no change,
Do you need more pure CPU power ? Just replace CPU card.
More memory ? Add it.
Hard disk or any permanent storage ? Add storage card.
etc., etc.

@TomjNorthIdaho you win the award for the longest post but it sounds like you want a x86 server.

Normis, thank you for asking for our opinions.

If it was at all possible, I would suggest Mikrotik make this new router with 2x modular "interface" bays. Users could then "cold swap" these modules to meet their specific requirements...

Making a modular router would:
- Allow various interface combinations, and would keep everyone happy
- Lower the manufacturing risk to Mikrotik (Bulk manufacture the base router, then JIT manufacture the interface modules)
- Allow users to change out the interface modules as their requirements change, e.g. they could start out with a 8x SFP+ module, then later on move to a QSP28 module
- Meet lower cost requirements, e.g. cost sensitive customer could buy base router + 1 PSU + 1 module card. Baller customer could buy base router + 2 PSU + 2 module cards.

Example interface modules could be:
1x QSFP28
2x QSFP28
4x SFP28
8-10x SFP+
16x SFP
16x 1gigabit Copper

If you do decide to go with "fixed" port configs though, my suggestions would be:

Spec1 For people who have exiting 10gigabit networks and do not need higher performance.

- 16x SFP+
- 1x Gigabit Ethernet

Spec2 For people wanting higher speed per "wavelength" or to future proof themselves.

- 8x SFP28
- 2x QSFP28
- 1x Gigabit Ethernet

This will allow 100gigabit throughput via the 8x SFP28 (4in, 4 out) or the 2x QSFP28 (1in, 1 out). It would also mean you could interconnect to other locations via redundant QSFP28 connections and then use SFP28 to connect customers, or to other equipment in the same DC.

I put the 1x Gigabit Ethernet on both of them, as this can be used to netinstall the routers, and for out-of-band management.

We are working on some future high speed routers with around 100Gbit throughput, and we would like your input on what types of ports you would want to see in such devices.
If you imagine a device with around 16 ports (just an example), how many of them would you want to be (just examples, you can add your own ideas):

SFP+ (10G)
SFP28 (25G)
QSFP (4x10G)
10Gbit copper (10G)
Management with PoE

Let us know in the replies here

Hello Normis
IMHO sfp+ on router more than enough.

Maybe not a specifica high speed router, but a PoE switch / router model which has configurable PoE voltages.

So each port should be able to set:

• no PoE

• passive PoE (and select voltage, mostly 24V)

• 48V / 56V PoE.

So that for example a Cisco or HP switch can be replaced by a Mikrotik.

The problem now is that if you have an existing Cisco / HP / Juniper etc.... PoE switch for xxx AP's, and you want to add Mikrotik, you cannot simple do that, since all professional AP's work on 48V. And most Mikrotik (and Ubiquity etc...) work on passive 24V. So you need to find local power (not always available) etc...

So I want to select the power voltage output on each port.

@TomjNorthIdaho you win the award for the longest post but it sounds like you want a x86 server.

I second the x86 architecture, but with the configuration and control of a routerboard

Port config, I'd like to see 16 or 24 ports SFP+ with an option on a couple of higher speed ports for interconnection.

Most importantly though, if its a rack model, it MUST have internal psu, with an optional secondary. No more damn wall warts in the rack!

Most importantly though, if its a rack model, it MUST have internal psu, with an optional secondary. No more damn wall warts in the rack!

This router sounds like the Next Gen CCR. The current CCR have internal PSU so I can't see Mikrotik changing now for no reason.

My wishes are for user replaceable PSU on the new product line, so that PSU can be replaced without removing the router from the rack.

And for -48volt DC input like the new RB1100AHx4 has this way they can be used in Telco co-lo rooms and road side cabinets that are almost always -48volt only.

I don't know which are the possibilities of the considered platform so cannot tailor port count/capacities to specific CPU lanes, but these would be sweet combos to have, I assume 1U and routing application (not switch):

- 12 SFP+, 3xQSFP
- 16 SFP+, 2xQSFP
- 16 SFP+, 4xQSFP
- 24 SFP+, 4xQSFP or 2xQSFP28 (2x100G)

The advantage of specifically manufactured hardware is that manufacture can test it at the specific configuration. If you take that away, is becomes the same as any other PC where you have no driver support.

Normis, every RM Router Should have a DC power option!


Sent from my iPhone using Tapatalk

48 port Gb/e and 4x 10Gb/e (SFP+) (TileGX based)
:
Or please support the following chipsets from Broadcom. The Trident II and the Tomahawk.
.

- Trident/Tomahawk is meant for switches not for router.

- Does TileGX has a future at Ezchip/Mellanox? There are other options: http://www.mellanox.com/page/npu_overview (NP/NPS)

I respectfully disagree. Those chipsets from Broadcom are fully functional and capable routers. I was hoping Mikrotik could jump into the bandwagon to disrupt Cisco/Juniper/Nokia and bring down the overall costs of networking. Eventually I'd like to start an ISP business, but well.....equipment isn't exactly cheap. That being said, it's not equipment that bothers me so much as having to grease the wheels by buying political power to get right of way and whatnot.

I respectfully disagree. Those chipsets from Broadcom are fully functional and capable routers. I was hoping Mikrotik could jump into the bandwagon to disrupt Cisco/Juniper/Nokia and bring down the overall costs of networking. Eventually I'd like to start an ISP business, but well.....equipment isn't exactly cheap. That being said, it's not equipment that bothers me so much as having to grease the wheels by buying political power to get right of way and whatnot.

The chipsets from Broadcom are being used in "routing switches" e.g. Brocade MLX/VDX, Nokia 7210 SAS-T/M and certain Arista switches.

They are great for fixed scenarios but are not good as a "General Purpose" router, and have some pretty hard limitations. e.g. Broadcom Trident II can do VXLAN and QinQ and vlan re-write, just not all together, and due to everything being "baked" into the silicone, that cannot be changed.

Mikrotik have a lot more experience now than when they started the CCR project, so I am confident they will select an architecture that is suitable for use as a general purpose router.

Normis, every RM Router Should have a DC power option!

every CCR should have a -48volt DC power option

If you says about ROUTER, this configuration will be interesting:
2 x Copper 1Gbit
10 x SFP+ 10Gbit
2 x Copper 10Gbit Ethernet
2 x QSFP 40Gbit

The chipsets from Broadcom are being used in "routing switches" e.g. Brocade MLX/VDX, Nokia 7210 SAS-T/M and certain Arista switches.

They are great for fixed scenarios but are not good as a "General Purpose" router, and have some pretty hard limitations. e.g. Broadcom Trident II can do VXLAN and QinQ and vlan re-write, just not all together, and due to everything being "baked" into the silicone, that cannot be changed.

Mikrotik have a lot more experience now than when they started the CCR project, so I am confident they will select an architecture that is suitable for use as a general purpose router.

Yes, they are layer 3 "routing switches". Basically routing functionality bolted on. The reason I was looking more at those is because that limitation that you're pointing out (which does exist) is one that GENERALLY isn't hit too often.I was more looking at the MPLS/VXLAN/QinQ/translational as features that can be done at line rate. Usually one would need one or the other, but not all at once. Now I am not pushing super hard to get those chipsets, I'm just more investigating and seeing if they are a fit. I personally would love to see line rate from Mikrotik. Or near line rate (which the Tile GX approach).

Something I was always hoping for was like a CRS226-24G-2S+RM chipset with the Tile GX for forwarding. Or CRS125-24G-1S-RM with Tile GX. Or even the smaller ones like the CRS210-8G-2S+IN.

In additional to as many high-speed ports as you can cram into the box either a line-card slot or just a bank of regular 1GB Ethernet ports can be extremely useful to designers and provide a high level of flexibility. Something like 6 - 12 x 1GB ports wouldn't hurt the back-plane of the box significantly I imagine. Something I hated with the earlier Cisco ISRs and even somewhat today is the difficulty to obtain the number of ports necessary to do anything other than trunk down to a L2 switch below it if you are doing anything more than a typical edge router with NAT. This is already a strength with most of your devices shipping with 4+ ports.

I respectfully disagree. Those chipsets from Broadcom are fully functional and capable routers. I was hoping Mikrotik could jump into the bandwagon to disrupt Cisco/Juniper/Nokia and bring down the overall costs of networking. Eventually I'd like to start an ISP business, but well.....equipment isn't exactly cheap. That being said, it's not equipment that bothers me so much as having to grease the wheels by buying political power to get right of way and whatnot.

The chipsets from Broadcom are being used in "routing switches" e.g. Brocade MLX/VDX, Nokia 7210 SAS-T/M and certain Arista switches.

They are great for fixed scenarios but are not good as a "General Purpose" router, and have some pretty hard limitations. e.g. Broadcom Trident II can do VXLAN and QinQ and vlan re-write, just not all together, and due to everything being "baked" into the silicone, that cannot be changed.

Mikrotik have a lot more experience now than when they started the CCR project, so I am confident they will select an architecture that is suitable for use as a general purpose router.

Great points as always NZ...

However, I will add that ASIC limitations are no longer an issue for any of the newer whitebox switch gear running the Barefoot Tofino chip as the ASIC can be programmed on the fly with the P4 language. Barefoot is being put in switches that would have otherwise run a Trident 2+ chipset.

However, I will add that ASIC limitations are no longer an issue for any of the newer whitebox switch gear running the Barefoot Tofino chip as the ASIC can be programmed on the fly with the P4 language. Barefoot is being put in switches that would have otherwise run a Trident 2+ chipset.

@IPANetEngineer I completely agree, the Barefoot Tofino is one of the most interesting developments in networking in the last 10 years.

It will be interesting to see what products make use of Tofino in the next 18 months.

I respectfully disagree. Those chipsets from Broadcom are fully functional and capable routers. I was hoping Mikrotik could jump into the bandwagon to disrupt Cisco/Juniper/Nokia and bring down the overall costs of networking. Eventually I'd like to start an ISP business, but well.....equipment isn't exactly cheap. That being said, it's not equipment that bothers me so much as having to grease the wheels by buying political power to get right of way and whatnot.

The chipsets from Broadcom are being used in "routing switches" e.g. Brocade MLX/VDX, Nokia 7210 SAS-T/M and certain Arista switches.

They are great for fixed scenarios but are not good as a "General Purpose" router, and have some pretty hard limitations. e.g. Broadcom Trident II can do VXLAN and QinQ and vlan re-write, just not all together, and due to everything being "baked" into the silicone, that cannot be changed.

Mikrotik have a lot more experience now than when they started the CCR project, so I am confident they will select an architecture that is suitable for use as a general purpose router.

Great points as always NZ...

However, I will add that ASIC limitations are no longer an issue for any of the newer whitebox switch gear running the Barefoot Tofino chip as the ASIC can be programmed on the fly with the P4 language. Barefoot is being put in switches that would have otherwise run a Trident 2+ chipset.

I assume that chip is a bit larger scale than Mikrotik is looking to go... given that it handles 65x the performance Mikrotik hinted towards in the initial post.

The question was on a router and around 100 Gbit throughput.

I'd think that would limit it basically to two QSFP-Ports (probably often used as "upstream"). I'd further propose 6x SFP+ - ports and additional 8x 10G copper (10GBASE-T) for compatibility with existing (even GBit) connections. I'd liked an extra (GBit) management lan port and serial connection, too.

It probably makes sense to have (at least optional, if need be external) redundant power supplies as well.

Even if reduced to 4x SFP+, 4x 10GBase-T, 8x 1000Base-T (maybe even partly or all with PoE+) plus mgmt-LAN & serial would it make a great cloud router switch ...

And go compete with china?

I think you should stick with wireless stuff, and maybe build a quality FDM full-duplex system with sub-1ms ping over 10km. We'd like that.

Coming from a Data Center background, I'd like to see:

Access / Distro #1
24x 1G / 10G Copper
4-6x SFP28 / QSFP28

Access / Distro #2
16x SFP / SFP+
4-6x SFP28 / QSFP28

Core
You could also run a "core" router with:
10x SFP28 / QSFP28
4x 1G / 10G Copper
or
4x SFP / SFP+

I'd still love to see you guys add VXLAN / EVPN features to your devices.

Let's not forget that QSFP and QSFP28 both support breakout or fanout cables.

Let's not forget that QSFP and QSFP28 both support breakout or fanout cables.

And Dont forget:

QSFP28 can do:
1-4 x 10
1-4 x 25
1-2 x 50
1 x 100

SFP28 can do:
1 x 10
1 x 25

QSFP+ can do:
1-4 x 10
1 x 40

SFP+ can do:
1 x 10
1 x 1

All speeds and breakout combinations should be supported on all the ports no shit as of now only 10 och sfp+ port

I'd still love to see you guys add VXLAN / EVPN features to your devices.

This is getting pretty OT now, but:

+1000

Please email support@mikrotik.com and lodge your official feature request. Eventually they will get annoyed enough by the constant emails and add it

I respectfully disagree. Those chipsets from Broadcom are fully functional and capable routers. I was hoping Mikrotik could jump into the bandwagon to disrupt Cisco/Juniper/Nokia and bring down the overall costs of networking. Eventually I'd like to start an ISP business, but well.....equipment isn't exactly cheap. That being said, it's not equipment that bothers me so much as having to grease the wheels by buying political power to get right of way and whatnot.

The chipsets from Broadcom are being used in "routing switches" e.g. Brocade MLX/VDX, Nokia 7210 SAS-T/M and certain Arista switches.

They are great for fixed scenarios but are not good as a "General Purpose" router, and have some pretty hard limitations. e.g. Broadcom Trident II can do VXLAN and QinQ and vlan re-write, just not all together, and due to everything being "baked" into the silicone, that cannot be changed.

Mikrotik have a lot more experience now than when they started the CCR project, so I am confident they will select an architecture that is suitable for use as a general purpose router.

Great points as always NZ...

However, I will add that ASIC limitations are no longer an issue for any of the newer whitebox switch gear running the Barefoot Tofino chip as the ASIC can be programmed on the fly with the P4 language. Barefoot is being put in switches that would have otherwise run a Trident 2+ chipset.

I assume that chip is a bit larger scale than Mikrotik is looking to go... given that it handles 65x the performance Mikrotik hinted towards in the initial post.

Mike, Barefoot is looking to scale down as well as up. It's a question I specifically asked them when I visited their corp HQ in January. Look for Barefoot going into smaller boxes in next 12 months.

I would like 8x QSFP+ in pretty much the same form factor/config as the CCR 1072.

That way I could do uplinks to my older model Force10 switches and backbone 40Gbps pretty cheap across longer distances.

LR QSFP+ modules are in the low hundreds range now. With 40km QSFP+ modules coming down in price as well.

I myself also talked about QSFP28, but as Normis mentioned, Mikrotik prepares for a 100Gbit/s router, not a terrabit one. Although I also want to build an inexpensive 100 Gbit/s core.

I would like to see more of a focus on the layer2 distribution side so it ties in with these new proposed routers. Otherwise we would need to pick other brands for L2 which we have been doing especially for SFP+ port density and POE. Stacking on the switches as well as the routers. More IPv6 support in ROS. No need for any LCD's. Removable dual power supplies.
Personally I do not need 10gb ethernet, but in the server/dc environment I can see the need.
Layer2 access:
24 SFP+ ports, dual PSU
24 SFP+ ports + 4 x QSFP, dual PSU
48 x 1Gbe + 2 x SFP+
24 x 10Gbe + 4 x QSFP
48 x 1Gbe POE at/af + 2 x SFP+

CCR's:
12 x SFP+ and 4 x 10Gbe combo ports
12 x SFP+ and 4 x QSFP

Please do a CLEAR separation between "enterprise" routers, and SOHO routers. I think it's time...

Enterprise Routers - PLEASE we are willing to pay, MAKE IT RELIABLE, make it PERFORM.

Interfaces, can be modular. 4 x 1GB, 4 x SPF, 2 x SFP+, 1 x QSFP+, etc... No need to have fixed ports. Same with power supplies, same with fans.

Just my 2c, but BIGGER routers means NOTHING when the CURRENT routers can't perform either. There's numerous flaws on your "flagship" CCR routers, which has little to no interest from MT to be fixed (PSUs, BGP, etc). I for one, won't even THINK about purchasing "bigger" routers from MT when the existing routers are so unreliable and under performing.

I know this is probably not going to go over well, but I'm going to say it anyway: it's time for Mikrotik to have an Apple equivalent of "Back to the Mac"; except the Mac is RouterOS. I don't want to see another piece of hardware. I want to see a commitment to releasing RouterOS 7. All these different devices they are releasing, yeah.. they make sense for a small portion of your customers. Why not focus some effort on something that _all_ your customers can benefit from? There are numerous wish lists on the forums talking about what we'd all like to see in v7, so it's not worth re-iterating here.

Hopefully someone out there will see the logic in what I'm saying.

I know this is probably not going to go over well, but I'm going to say it anyway: it's time for Mikrotik to have an Apple equivalent of "Back to the Mac"; except the Mac is RouterOS. I don't want to see another piece of hardware. I want to see a commitment to releasing RouterOS 7. All these different devices they are releasing, yeah.. they make sense for a small portion of your customers. Why not focus some effort on something that _all_ your customers can benefit from? There are numerous wish lists on the forums talking about what we'd all like to see in v7, so it's not worth re-iterating here.

Hopefully someone out there will see the logic in what I'm saying.

I've said this many times and I get yelled at and disparaged.

ROS7 > *

ROS7 needs to come out before 'tik is able to really make more inroads.

I know this is probably not going to go over well, but I'm going to say it anyway: it's time for Mikrotik to have an Apple equivalent of "Back to the Mac"; except the Mac is RouterOS. I don't want to see another piece of hardware. I want to see a commitment to releasing RouterOS 7. All these different devices they are releasing, yeah.. they make sense for a small portion of your customers. Why not focus some effort on something that _all_ your customers can benefit from? There are numerous wish lists on the forums talking about what we'd all like to see in v7, so it's not worth re-iterating here.

Hopefully someone out there will see the logic in what I'm saying.

Software engineers != Hardware Engineers

Not that I disagree with the importance of getting v7 done and done properly.

I know this is probably not going to go over well, but I'm going to say it anyway: it's time for Mikrotik to have an Apple equivalent of "Back to the Mac"; except the Mac is RouterOS. I don't want to see another piece of hardware. I want to see a commitment to releasing RouterOS 7. All these different devices they are releasing, yeah.. they make sense for a small portion of your customers. Why not focus some effort on something that _all_ your customers can benefit from? There are numerous wish lists on the forums talking about what we'd all like to see in v7, so it's not worth re-iterating here.

Hopefully someone out there will see the logic in what I'm saying.

Software engineers != Hardware Engineers

Well there's also no point in bringing out hardware, if the software to utilize the hardware, isn't there

Just different requirements\designs. The CCRs are great performance for the price when you're not considering large routing tables. Whatever this platform is may have big enough cores to overcome the large routing table issue. BGP works just fine on x86 and CHR. If not, you use these boxes for MPLS and they don't carry full Internet tables. Just stick on an x86 box for your full tables routers.

Let's keep this thread on topic. If you must whine, elegantly incorporate it in an appropriate response to the original request.

Just different requirements\designs. The CCRs are great performance for the price when you're not considering large routing tables. Whatever this platform is may have big enough cores to overcome the large routing table issue. BGP works just fine on x86 and CHR. If not, you use these boxes for MPLS and they don't carry full Internet tables. Just stick on an x86 box for your full tables routers.

Let's keep this thread on topic. If you must whine, elegantly incorporate it in an appropriate response to the original request.

x86 which is being phased out (and supports a few, at best hardware devices. MT even said no more new drives till the mythical v7 makes an appearance), CHR which to say the least carries piles of additional licensing fees, and misses crucial functionality. Hell, I may just as well take a trusty linux distro with Quagga, Bird or something else then I suppose. Why then bother with MT, and MT's hardware at all (and what about the thousands I've already invested in hardware that promised the world and delivered everything but)? I suppose it's much more important to rather have a ROUTER running a SMB server, or HTTP Proxy server, or even insecure DNS servers by default

Anyways, that's pretty much the response I expected yes. Say anything bad and you are "whining" Never allowed any criticism against MT, they can't do anything wrong... All I'm saying, MT needs to get its priorities straight... There's a lot of people starting to notice these... issues...

Anyways, that's pretty much the response I expected yes. Say anything bad and you are "whining" Never allowed any criticism against MT, they can't do anything wrong... All I'm saying, MT needs to get its priorities straight... There's a lot of people starting to notice these... issues...

If you don't think I criticize or allow criticism, you clearly haven't been paying attention.

Look guys, I'm merely trying to point out that MT's customers would be better served by concentrating their efforts on RouterOS v7 instead of releasing all these little "trinket" devices. Has anyone else noted that the newsletters are mostly full of hardware announcements? I get that hardware engineers != software engineers. That actually makes my point more strongly: it's like MT forgot that RouterOS is the entire basis of their platform. Really, I think the argument could be made that the hardware doesn't really matter all that much. RouterOS can run on almost anything these days.

it's time for Mikrotik to have an Apple equivalent of "Back to the Mac"; except the Mac is RouterOS.

Make RouterOS great again ?

Take into account only a few devices have 2 power supply built-in, I really doubt you'll do well on really high-end market. Cheap China-grade external "hey ma, I got my phone power supply for my router!" typed power sources are something that won't look nice for anything high-end.

Who'll pay for 100G's while unsure in device just for its low-priced components? If so, then high-availability features must be built-in deeply into ROS. Stability and failover is the key, and MT devices used to hang sometime so far, this is blocker for high-load high-end market, too.

What you may want to do is to create 10G switches (with ASIC and ROS in it and a lot of memory, not just SwicthOS), this may be useful for poor-man clusters and some low-end datacenters, but hight-end is something different...

None of the rack mountable CCR units has an external power supply

it's time for Mikrotik to have an Apple equivalent of "Back to the Mac"; except the Mac is RouterOS.

Make RouterOS great again ?

omg you savage! =)

None of the rack mountable CCR units has an external power supply

Ok, then lets's count how many of them has two PSU? I suspect there is only ONE (ok, two, since one was a replacemnet for another one) When it comes to this point all I got from Mikrotik was advice to supply second power source via PoE, either via MT-compatible switch (tricky thing to find to use in rack in high-end DC) or via use PoE injector with - yes - China made cheap external PSU.

So you're still go to high-end market?

Ok, I understand: your CCRs are so strong and reliable so you don't even believe anyone may want to have two PSU? Then that's good approach, but guys at huge DCs are so used to see backup PSU everywhere so they won't get that for real.

So you may not worry for ports. Start with PSU and overall durability. Looks like your newsletters are not at all monthly, and you still have no service contracts, so it looks pretty unserious for someone who try to deal with 100G ports.

Ok, I understand: your CCRs are so strong and reliable so you don't even believe anyone may want to have two PSU?

That's not even the question of how reliable the PSUs themselves are. That's the question of the ability to power the router from two separate UPSs and be able to keep the router running while you're replacing the batteries in one of them.

No one argues that having two PSUs in a high-end device is a must. However this topic is about something else, right? So, please, guys, lets stay on topic here!

None of the rack mountable CCR units has an external power supply

But I would like a desktop unit with a single internal PSU.

None of the rack mountable CCR units has an external power supply

But I would like a desktop unit with a single internal PSU.

Not a problem, but this is not high-end device, right? And this topic is for 100G ports, something you just don't need in desktop unit, after all.

We are working on some future high speed routers with around 100Gbit throughput,

So MikroTik is now working on a 100 Gbit/s router. The only decision is which ports that device should support. Can we get a hint what ASIC / platform is currently used?

For first high end routers i think better is create a product with most used ports. Due to dropping of 100Ge ports price, most datacenters are replacing SFP28 QSFP etc with them.

Modular router is fundamental requirement
I think somethinf like:

- Default:
1 x Fe Mgmt Port
8 x 10SFP+ port

- Free module Slot where customers can choose between
4 x 10SFP+ port
8 x 10SFP+ port
n x 100Gbps Port

Another requirement is to optimize the handle of bonding ports that now impact too much the cpu

Modular router is fundamental requirement

Wish to look at it! Modular 13U router from Mikrotik with add-in blades full of different ports, with built-in support for hardware redundancy of CPU cards, storage.

And with (still) one internal PSU

Really, I suspect a lot of people would like to see not 100G, but simple to say modular (or at least stackable) solutions from MT. It you build decent LAN you'll need at least 100-150 client LAN ports per closet and you have to put switches each of 24 or 52 1G ports then connect it to central unit with 10G SFP+ but - no redundancy!

So to say, I'd ask for:

- central switch (capable of hardware stack, so I can put two of them) full of 10G, that'll have at least 2x10G or more to uplink or to interconnect to different closet in different room.
- edge layer switch capable to connect with two 10G links to these "central" switches
all of these should offer full stack of switch technologies, and good diagnostics/manageability.

To my surprise, even when we'd like to feel we have a lot of traffic in the network in fact not always we need 100G ports. What we really need is some approach to build robust self-healing system so big network in office or in datacenter won't be affected by single device fault. So modular approach is a good thing, indeed.

By the way, modular devices are more controllable and power-efficient, so for say 10 line cards worth of 50 ports each you can use 2 or 4 PSU per chassis, not 10 (one for each small switch) and the overall robustness will be higher. Remote management will be much better when you can control the whole chassis with cards in it due, etc.

So, please, aim at modular and longer-live devices, not at cheaper "save on everything" ones!

You''re confusing "modular" word. Modular not mean 13u or bigger router. For example ASR 1001 is a modular router and is 1u.



You can setup a 1u or 2u router with fixed ports and a module for expansion to save $$ if you need expansion

Again, remember that with big ports, you can do breakout\fanout cables to put multiple smaller ports into a bigger port. That way you can have 40G or 100G ports, but can also do 10G or 25G off of them.

I would prefer modular port cards.

for example the box should have the rj45 mgt port, a serial one and two sfp+ combo ports by default

everything else should be on the extensions cards.

you need another 8 copper ports? buy a card. qsfp for the router on a stick configuration? select the right extension card.

the cpu/chasis should only have two versions:
a full one (all cores whatever) and a reasonable one.

right now it is really painfull to select the right ccr.

one of the extension cards could use infiniband to expand to another next gen ccr to extend the ports.

and please do not forget hotswapable psu.

A modular chassis is something I'd accept, but I'd assume it's a lot more expensive than a fixed switch.

Yes, redundant hot-swap power supplies are a must. I like the power layout of the RB1100AHx4.

When I mentioned modularity I was thinking of the Cisco ASR 1001/1002, Huawei NE05E-S2 and at the higher end the Juniper MX104. Not some 10+ RU monstrosity.

The line cards in a 1 or 2 RU router could be made cost effectively by just containing the appropriate SERDES to PHY conversions.

If you bring out ROS7 (BGP Multicore), then please build routers with 2-4 Tilera72 or newer CPU's with:
4x CFP4 (100GE) (router should be able to handle a total amount of 200 Gbit/s)
8x SFP+
1x RJ45 Gbit

I think the future is CFP4 at the moment.

This is the market that Mikrotik does not reach at this time (Brocade CES, Juniper, Cisco,...).

But why not build a plattform like Brocade MLX with linecards (24x RJ45, 8x SFP+, 4x CFP4,...) and CPU cards that can be placed?
A case with 2 power supplies and line card slots could be cheap to produce.

If you bring out ROS7 (BGP Multicore), then please build routers with 2-4 Tilera72 or newer CPU's with:
4x CFP4 (100GE) (router should be able to handle a total amount of 200 Gbit/s)
8x SFP+
1x RJ45 Gbit

I think the future is CFP4 at the moment.

This is the market that Mikrotik does not reach at this time (Brocade CES, Juniper, Cisco,...).

But why not build a plattform like Brocade MLX with linecards (24x RJ45, 8x SFP+, 4x CFP4,...) and CPU cards that can be placed?
A case with 2 power supplies and line card slots could be cheap to produce.

Will people please stop talking about multicore BGP! It isn't going to happen! The performance enhancements are coming from other areas, not multi thread support.

Will people please stop talking about multicore BGP! It isn't going to happen! The performance enhancements are coming from other areas, not multi thread support.

Well, they did do some work on it and were talking about it. I personally am not a big fan of making it too complicated....I think 1 thread for 1 neighbor should be good enough. But that's up to 'tik. However their competitors are multi-threading all of their daemons so........there's that.

If Mikrotik is to keep up then they kinda have to do so.

I personally have never had that much of a problem with the "speed" of a router/network convergence. This coming from working on networks that throw around tens of Tb/sec, the speed of convergence is generally not that much of a problem. Unless one works in like financial services I don't see this as much of a big deal.

To be honest, having separate daemons (as in a separate daemon for LDP, BGP, OSPF, ISIS, RSVP, so on) is more important than the daemons being multi-threaded themselves.

It's all about what Mikrotik is focusing on. If they want to to compete with Cisco/Juniper then they will need to get their software in gear. If they don't then they shouldn't. I personally would love to have an equivalent to Cisco/Juniper. Currently, 'tik is just not there. They do ok, but they aren't there yet.

Will people please stop talking about multicore BGP! It isn't going to happen! The performance enhancements are coming from other areas, not multi thread support.

Well, they did do some work on it and were talking about it. I personally am not a big fan of making it too complicated....I think 1 thread for 1 neighbor should be good enough. But that's up to 'tik. However their competitors are multi-threading all of their daemons so........there's that.

If Mikrotik is to keep up then they kinda have to do so.

I personally have never had that much of a problem with the "speed" of a router/network convergence. This coming from working on networks that throw around tens of Tb/sec, the speed of convergence is generally not that much of a problem. Unless one works in like financial services I don't see this as much of a big deal.

To be honest, having separate daemons (as in a separate daemon for LDP, BGP, OSPF, ISIS, RSVP, so on) is more important than the daemons being multithreaded themselves.

Who is doing multi-threaded BGP? I think there's one or two platforms out of all BGP platforms that are multithreaded. Gah, I got suckered into a hijack....

To attempt to bring it back on-topic, bigger CPU cores postpones the need for v7's new routing. Unless MT is going x86 or CHR with the new platform, version 6 will need major rewrites, which just seems silly at this point.

To attempt to bring it back on-topic, bigger CPU cores postpones the need for v7's new routing. Unless MT is going x86 or CHR with the new platform, version 6 will need major rewrites, which just seems silly at this point.

At one point in time I thought there was a powerpoint presentation and a video talking about multithreaded BGP daemon for CCRs in v7. I'm having a hard time finding it now but I did read through it a few times.

Hijack wasn't meant at least on my end.

On the hardware front, the reason I personally wanted to go towards the Broadcom or Octeon chipsets is mainly because one thing that networks need is wire rate (or near wire rate) performance for all packet types. I personally as a network engineer would be fine building a network that can wire rate 256 byte packets, and really any smaller than that requires specific ASICs even. Most network vendors struggle to do wire rate at 64 byte packets. At the moment the 'tik can't really do near wire rate when using non-fast path/firewall filters/NAT/marking/QoS/MPLS edge services. To build the networks that are needed for today one kinda needs to be able to approach wire rate even with services enabled. Either 'tik needs to do multiple Tilera chips working together to provide the processing power needed or they need to go the ASIC route. Either would be fine with me. Or they can stay where they are, but they shouldn't expect to be able to break into the network that will buy tens of thousands of 'tik routers without those features. I would personally love to be able to use them as MPLS services edge boxes but, they just aren't there yet. Would love to use them as core devices but they aren't there yet either.

To attempt to bring it back on-topic, bigger CPU cores postpones the need for v7's new routing. Unless MT is going x86 or CHR with the new platform, version 6 will need major rewrites, which just seems silly at this point.

At one point in time I thought there was a powerpoint presentation and a video talking about multithreaded BGP daemon for CCRs in v7. I'm having a hard time finding it now but I did read through it a few times.

Hijack wasn't meant at least on my end.

On the hardware front, the reason I personally wanted to go towards the Broadcom or Octeon chipsets is mainly because one thing that networks need is wire rate (or near wire rate) performance for all packet types. I personally as a network engineer would be fine building a network that can wire rate 256 byte packets, and really any smaller than that requires specific ASICs even. Most network vendors struggle to do wire rate at 64 byte packets. At the moment the 'tik can't really do near wire rate when using non-fast path/firewall filters/NAT/marking/QoS/MPLS edge services. To build the networks that are needed for today one kinda needs to be able to approach wire rate even with services enabled. Either 'tik needs to do multiple Tilera chips working together to provide the processing power needed or they need to go the ASIC route. Either would be fine with me. Or they can stay where they are, but they shouldn't expect to be able to break into the network that will buy tens of thousands of 'tik routers without those features. I would personally love to be able to use them as MPLS services edge boxes but, they just aren't there yet. Would love to use them as core devices but they aren't there yet either.

https://youtu.be/ihZiAC-Rox8?t=37m09s
https://mum.mikrotik.com//presentations/US14/ccrros.pdf (page 32)

Octeon would be a good path for sure.

i think 2.5g ethernet and 5g ethernet must be take in consideration (for smaller routers off course) maybe as an intermediate upgrade without going to full 10g ethernet over copper 6a category

Now I am at 10G network, and I prefer to deploy the 10G to 25G to 100G migration path instead of the 40G to 100G, so I hope the high speed routers with 25G ports.
The single lane 25G network will become more popular in the near future.

Now I am at 10G network, and I prefer to deploy the 10G to 25G to 100G migration path instead of the 40G to 100G, so I hope the high speed routers with 25G ports.
The single lane 25G network will become more popular in the near future.

Agreed.

We are already seeing 25G switches from the major datacentre switch vendors.

Now I am at 10G network, and I prefer to deploy the 10G to 25G to 100G migration path instead of the 40G to 100G, so I hope the high speed routers with 25G ports.
The single lane 25G network will become more popular in the near future.

In my recommendation, you could do both.

I would like to see CCR with
2x SFP + and 8 SFP
You don't have anything that has 2xSFP+ ports and several SFP ports (Not counting the CCR1072-1G-8S + which is terribly expensive)

At the moment, I need 2xSFP + and 5xSFP, so I use 2x CCR + bonding on ethernet ports - I dont like this solution...

I'l vote for SFP28 - 25Gbit should be future, but few people know about it, sadly.

It was mentioned on AWS:reinvent
"The custom chips also power a custom AWS network architecture that uses 25Gb Ethernet, a format Amazon believes is actually more scalable and efficient than the 10Gb and 40Gb Ethernet standards commonly used."

http://www.plexxi.com/2014/07/whats-25- ... rnet-want/

And for smaller boxes (home usage?) 5Gbit makes sense, cause it can reuse cat5e cables

http://www.anandtech.com/show/11368/aqu ... 25g1g-nics

4 40 gig QSFP
8-10 10 gig RJ45
8-10 10 gig SFP+

As long as the QSFP ports can be configured with LACP, this is a pretty flexible setup. Breakout cables were mentioned, they can be used if needed to match LC infrastructure.

10 gig RJ45 is not super duper popular, but it's flexible for older devices since it will negotiate all the way down to 10baseT. And I really don't like using RJ45 SFPs. I think a 10 gig one is not cheap and not common.

I never understood the argument for putting fewer ports on a router. Just shut the ports not in use, and don't use more than bandwidth will allow for. Simple right?

based on our own needs here, i'd like to especially have more ports in a single router (given the CPU can handle it).

example 1 (medium thoughput, actual router, not CRS):
24 SFP
2 SFP+

example 2 (high throughput):
8 SFP
8 SFP+

example 3 (very high throughput):
8 SFP+
4 QSFP

in all three cases, both PoE-in and PoE-out would be nice if possible.

These routers are for distribution purpose only?

Like not for home use right? I am bit confuse. I have a 100mbps connection and Gbit is a lot considering the experience I get on 100.

Like not for home use right? I am bit confuse. I have a 100mbps connection and Gbit is a lot considering the experience I get on 100.

Yes, these are for a new tier of high end Mikrotik routers. As they said initially, 100 gigabit capacity.

I would like to see a modular router with add in cards for 25G, 40G, and possibly 100G. If I buy a router that I am doing 25G today I want to be able to upgrade that to 100G at one point.

48x SFP+
6x QSFP

as a BGP router...


edit: any ETA?

48x SFP+
6x QSFP

as a BGP router...


edit: any ETA?

+1

Hardware like
https://whiteboxswitch.com/products/qua ... -with-onie
http://www.edge-core.com/productsInfo.p ... =59&id=119

have very good performance/price ratio, but it is still too complicated to setup and maintain. With RouterOS this should be (and will be) killer.

48x SFP+
6x QSFP

as a BGP router...


edit: any ETA?

I think Mikrotik are talking about a high-touch(services) router, not a low-touch router. This will be a services router like the CCR that can run PPP/NAT type services, not a "low-touch" router/switch that can just forward frames quickly.

why even people post switch gear links? we are talking about the next gen ccr

I want 100gb and -48vdc power.

I want 100gb and -48vdc power.

Don't we all ?

I think Mikrotik are talking about a high-touch(services) router, not a low-touch router. This will be a services router like the CCR that can run PPP/NAT type services, not a "low-touch" router/switch that can just forward frames quickly.

for BGP in IX we need something which have as much ports as possible. And quick forwarding/routing is they key functionality
inspiration here .. http://e.huawei.com/en/products/enterpr ... s/ne/ne40e

why even people post switch gear links? we are talking about the next gen ccr

now and in the future, there is not difference between L3 switch/router, speed is the key... hardware ASIC with TCAM is what you need.

Like not for home use right? I am bit confuse. I have a 100mbps connection and Gbit is a lot considering the experience I get on 100.

Yes, these are for a new tier of high end Mikrotik routers. As they said initially, 100 gigabit capacity.

Well this is good news then! So can I get it on my home network?

Also, I would like to know if these routers will allow me to distribute Internet bandwidth separately to other connected devices?

48x SFP+
6x QSFP

as a BGP router...


edit: any ETA?

I think Mikrotik are talking about a high-touch(services) router, not a low-touch router. This will be a services router like the CCR that can run PPP/NAT type services, not a "low-touch" router/switch that can just forward frames quickly.

for BGP in IX we need something which have as much ports as possible. And quick forwarding/routing is they key functionality
inspiration here .. http://e.huawei.com/en/products/enterpr ... s/ne/ne40e

An IX shouldn't be doing any routing. An IX should only be switching. Layer 2 only.

For the "more ports" people, check out 802.1BR

If Mikrotik were to implement 802.1BR on their routers and switches, it would allow "port extension", e.g. there could be a Mikrotik Router connected to a Mikrotik Switch in "Port Extender" mode, all of the ports on the switch would then appear as if they were local ports on the router.

802.1BR is supported by the switch chips used in the new CRS3xx series of switches, so it would be possible for Mikrotik to add the ability to use the CRS3xx as port extenders for their routers.

For the "more ports" people, check out 802.1BR

If Mikrotik were to implement 802.1BR on their routers and switches, it would allow "port extension", e.g. there could be a Mikrotik Router connected to a Mikrotik Switch in "Port Extender" mode, all of the ports on the switch would then appear as if they were local ports on the router.

until then 802.1ad can be used to connect CRS switches (CRS317-1G-16S+ will make lots of things possible) to CCRs as port extenders

there was great a presentation about this solution last year on Budapest MUM. unfortunately it was in hungarian, that's why the presentation is not available in pdf

An IX shouldn't be doing any routing. An IX should only be switching. Layer 2 only.

BGP is L2 ?

For the "more ports" people, check out 802.1BR

If Mikrotik were to implement 802.1BR on their routers and switches, it would allow "port extension", e.g. there could be a Mikrotik Router connected to a Mikrotik Switch in "Port Extender" mode, all of the ports on the switch would then appear as if they were local ports on the router.

802.1BR is supported by the switch chips used in the new CRS3xx series of switches, so it would be possible for Mikrotik to add the ability to use the CRS3xx as port extenders for their routers.

why two or more boxes? it is twice more chance something went broken. One box with enough SFP+ and QSFP port is preferred

48x SFP+
6x QSFP

as a BGP router...


edit: any ETA?

I think Mikrotik are talking about a high-touch(services) router, not a low-touch router. This will be a services router like the CCR that can run PPP/NAT type services, not a "low-touch" router/switch that can just forward frames quickly.

for BGP in IX we need something which have as much ports as possible. And quick forwarding/routing is they key functionality
inspiration here .. http://e.huawei.com/en/products/enterpr ... s/ne/ne40e

An IX shouldn't be doing any routing. An IX should only be switching. Layer 2 only.

BGP is L2 ?

The IX infrastructure doesn't do BGP. It just does basic Ethernet. Many (most) do have route servers that do run BGP, but those are best ran on servers, not routers.

The IX infrastructure doesn't do BGP. It just does basic Ethernet. Many (most) do have route servers that do run BGP, but those are best ran on servers, not routers.

Brocade, Juniper, Arista, Cisco, Huawei ... do you really mean "servers" ? Sorry maybe we talking about two other things, in IXP we have routers and interconnection with other providers by BGP, public IX have "routing servers" most of them Bird I guess.
But if you want to peer with other providers, you need router, with many ports, some ports you use for interconnect, some for transport. Ok, you can do it by two boxes, switch + router, but why? One box is better in may ways. Power, failure, security, backup ....

Could you be so kind and do not cite full post if you answer it ? If you need emphasize some part of previous post just cite this part.

BGP is L2 ?

The IX infrastructure doesn't do BGP. It just does basic Ethernet. Many (most) do have route servers that do run BGP, but those are best ran on servers, not routers.

Brocade, Juniper, Arista, Cisco, Huawei ... do you really mean "servers" ? Sorry maybe we talking about two other things, in IXP we have routers and interconnection with other providers by BGP, public IX have "routing servers" most of them Bird I guess.
But if you want to peer with other providers, you need router, with many ports, some ports you use for interconnect, some for transport. Ok, you can do it by two boxes, switch + router, but why? One box is better in may ways. Power, failure, security, backup ....

I have built and operate two Internet exchanges and am in regular discussion with my colleagues on IX operations. If you are running an IX, you want a switch, not a router. This is because Internet exchanges don't route themselves. They provide a platform to allow operators to connect to. Those operators then run BGP.

Yes, I really do mean servers and most IXes do run their route servers on BIRD, but OpenBGPd is making a comeback. We have built ours on OpenBGPd and are loving it.

If you're an operator connecting to an IX, you only need one port to connect to everybody else. That's the point of an Internet exchange. Obviously you would need other ports for transport, transit, etc. but they have nothing to do with the IX.

https://en.wikipedia.org/wiki/Internet_exchange_point

However, I feel as though we have gone far beyond the OP's question in explaining to you how an IX works. We can continue this conversation elsewhere, if you'd like.

Hammy,
Sorry but I have to ask as moderator: Can you read ? Do you understand what I am asking for ?
I have just asked not to cite post under post ... yes, yes, I know that your latest one is not strictly citation "post under post".

PLEASE edit your post and shorten it.

https://en.wikipedia.org/wiki/Internet_exchange_point

However, I feel as though we have gone far beyond the OP's question in explaining to you how an IX works. We can continue this conversation elsewhere, if you'd like.

no flame, but we prefer private direct peering (dedicated fiber), not through public IX. So from our point of view, as many ports as router can have, is better for us.

I would like a a 48 SFP's with one or 2 SFP+'s
ideally if it can be Stack-able to have expansion

I would like a a 48 SFP's with one or 2 SFP+'s
ideally if it can be Stack-able to have expansion

That sounds like a switch rather than a router..

None of the rack mountable CCR units has an external power supply

But I would like a desktop unit with a single internal PSU.

Not a problem, but this is not high-end device, right? And this topic is for 100G ports, something you just don't need in desktop unit, after all.

This topic is not about 100Gb ports, but about 100Gb throughput.

None of the rack mountable CCR units has an external power supply

But I would like a desktop unit with a single internal PSU.

Not a problem, but this is not high-end device, right? And this topic is for 100G ports, something you just don't need in desktop unit, after all.

This topic is not about 100Gb ports, but about 100Gb throughput.

Which you also don't typically need in a desktop unit. Most of those commenting seem to have lost the initial scope of the request.

Way, way off topic, sorry:

but OpenBGPd is making a comeback. We have built ours on OpenBGPd and are loving it.

What OS and what hardware?

I played with it on a PC just to see how it worked but it's been a while. Would be cool to deploy that on OpenBSD on beefy hardware, should be solid based on what I've seen.

Way, way off topic, sorry:

but OpenBGPd is making a comeback. We have built ours on OpenBGPd and are loving it.

What OS and what hardware?

I played with it on a PC just to see how it worked but it's been a while. Would be cool to deploy that on OpenBSD on beefy hardware, should be solid based on what I've seen.

Running OpenBSD on Proxmox and vSphere VMs now. Well, some older Dell hardware too. Low load, so don't need much power behind it.

I think I will add another 2 cents of my thoughts ...

I would like to see some copper 10-gig ports. (or 10/100/1-gig/10-gig ports).
How about something where we have 4 SFP+ ports and all Ethernet ports are 10/100/1-gig/10-gig ports and all ports are on a switch-chip.

My reasoning for the copper 10-Gig Ethernet ports is that servers are now available with copper 10-gig Ethernet ports.

North Idaho Tom Jones

Bump, @Mikrotik can you give any details of what is potentially being planned? Are you going to be moving of the Tilera platform? Marvell? Broadcom? 2018 launch? Faster Mhz CPU's?

We only gather ideas, but there is nothing to announce yet. When we have something ready, we will announce it in our MUM conference.

@normis for me new router need to be able to handle medium/high pps during attacks. We have lot of customers that joined our DDoS protected transit service because they have 100% CPU usage and loose router access during tcp floods (ccr1036 and ccr1072 devices).

As personal experience, on our CCR1072 we're able to push 15-20Gbps on exchanges without any issue. .Is useless create >100Gbps routers capacity when 1gbps tcp can block him (like ccr1072 now).

normis, we need sfp28 and qfp28

QSFP28

Have to step in here...

the next logical step would be a SFP28 / QSFP uplink SWITCH - routing not SO much needed, you can start with an Ethernet switch (though RouterOS would be appreciated). "Cloud Core" is moving to 25gbit, not alone because Intel has no network card (that I could find) in the 10g range that supports RDMA. Shame. And the Mellanox for a similar price to the Intel 710 - are SFP28. So, next round of servers here gets 25g Ethernet.

Cheapest switch I could find is https://www.fs.com/products/69375.html - 48 SFP28 and 6 QSFP for uplink. Which is too large for MY needs - a 16+2 port would be enough. Still, that things runs at 5000USD. WIthout OS - which adds significantly. On the plus side it is an ONIE switch, so you can choose the OS to use (hint: Mikrotik ONIE would be nice, handles flexible hardware). There is one free from MS, but it is db driven (no config command line, pulls config from a database - VERY nice for a large setup, they use that on Azure, but very limited on smaller scale setups). There is a smalle rversion (18xSFP28, 4xQSFP28) that costs around 6000 complete with OS.

We are working on some future high speed routers with around 100Gbit throughput, and we would like your input on what types of ports you would want to see in such devices.
If you imagine a device with around 16 ports (just an example), how many of them would you want to be (just examples, you can add your own ideas):

SFP+ (10G)
SFP28 (25G)
QSFP (4x10G)
10Gbit copper (10G)
Management with PoE

Let us know in the replies here

That is great, I hope you can launch a router with at least 4 QSPF ports.

We are working on some future high speed routers with around 100Gbit throughput, and we would like your input on what types of ports you would want to see in such devices.
If you imagine a device with around 16 ports (just an example), how many of them would you want to be (just examples, you can add your own ideas):

SFP+ (10G)
SFP28 (25G)
QSFP (4x10G)
10Gbit copper (10G)
Management with PoE

Let us know in the replies here

Question re "...routers with around 100Gbit throughput..."
To better understand the question and better formulate a response, I would like to ask what type/brand of CPU(s) are being considered to acheive 100Gbit routed L3 throughput ?

North Idaho Tom Jones

I would be amazing to have this Mikrotik switch:
24 SFP
4SFP+ (10G)
Two gigabit copper ports
We want to upgrade our network and replace Alcatel switches with Mikrotk.

12 x SFP+ (10G)
4 x QSFP (4x10G)
1 x Management (no PoE)
1 x serial management

second that.

For example,

I working now on a project for vip-class residential high-rise buildings. And we planning to make 10GFTTH. What we want to have on one site:

1. Stackable switch (Layer 2 and Layer 3 switching) for 500 SFP+ ports (in total) and 10x100Gb ports for uplink, with 128 Gb/s cross-connection between switches (connecting as ring).

2. For apartments we need user router with 2 (or 4) SFP+, 6 ports 10GBASE-T, and 10 ports GBASE-T (maybe few RS485 for telemetry, IoT, alarms).

3. Something like UBNT UAP-XG (bgn/ac/Wave2, if it will also have /ad this will be good) with real seamless WiFi roaming.

This APs and router will installed in apartments as defaults.

it has been more than one years, still don't see 100G port router from mikrotik, any new update?

Any news about this one? Any hints, any signs? Any rumors?
Back to CPUs/ASICs: Cavium´s Thunder X2 or Octeon TX look great.

=> list price for 32 core Cavium´s CN9980 is $1795

What's the status on these new router models?

We are working on some future high speed routers with around 100Gbit throughput, and we would like your input on what types of ports you would want to see in such devices.
If you imagine a device with around 16 ports (just an example), how many of them would you want to be (just examples, you can add your own ideas):

SFP+ (10G)
SFP28 (25G)
QSFP (4x10G)
10Gbit copper (10G)
Management with PoE

Let us know in the replies here

6x10Gbit Copper, 4xSFP+, 2xSFP28

The 40G we use still requires LC.

If 40G switching is already good for CPE, I think core would need 100G.

40gig is older generation now. No new gear is coming out with 40G on it any more, everything new is 100G QSFP28 as the chipsets cost the same whether they have 40G or 100G ports.

Also, the carrier/service provider industry has never embraced 40G, it has always been 10G, 25G, 100G, 200G and now 400G.

You could buy a small Ryzen NAS and a couple 40G cards for your home. You cannot do that with 100G.

I did not say 40G for the datacenter.

If you can plug a 40G cable into the 100G port and a small router/switch does not become much more expensive, 100G would be preferred for such a device, as one SSD is enough to saturate 40G. But that is not the subject of the discussion.

Hi!

I think you first need to learn how to make stackable hardware.

WBR,
Fyodor.

Hi!

I think you first need to learn how to make stackable hardware.

WBR,
Fyodor.

Which other routers are stackable ?

Good day!
Here is my idea

1) The main thing is that the router should have power supply modules like on CCR1072 (48v + 48v, 48v + 220v, 220v + 220v)
2) The router should have a minimum of 8x10GB (SFP +) and a 2x40GB interface. And so, you can create a router with 8x10GB (SFP +) and a slot for modules with interfaces 2x25GB or 2x40GB or 2x25GB + 2x40GB interface.
3) A CPU that will, without any problems, handle this interface volume on the switching ~ 320/480/960GB router. Because with this amount of data, many will use Mikrotik CCRxxxx as a core router with BGP Full View.
4) You can increase the size of the router by 2U or 3U. Not so much difference between 1U or 2U or 3U.

Best Regards.

Given that there are switch chips and encryption chip, how about other types of specialized ASICs to further offload the cpu and give the boxes better performance.

A firewall chip perhaps??


Sent from my cell phone. Sorry for the errors.

Given that there are switch chips and encryption chip, how about other types of specialized ASICs to further offload the cpu and give the boxes better performance.

A firewall chip perhaps??


Sent from my cell phone. Sorry for the errors.

Switch chips typically have ACL capabilities built-in.

Given that there are switch chips and encryption chip, how about other types of specialized ASICs to further offload the cpu and give the boxes better performance.

A firewall chip perhaps??


Sent from my cell phone. Sorry for the errors.

Switch chips typically have ACL capabilities built-in.

Limited to layer 2 I assume??

Sent from my cell phone. Sorry for the errors.

When it comes to future new features on new designs of Mikrotik devices ,...

I would like to propose:
#1: CRS (Bad Boy)
- An Intel XEON processor on a motherboard (fast clock with lots of CPU cache & ability to disable CPU hyper-threading)
- Zero 1-Gig ports (no 1-Gig only ports)
- Qty ten SFP+ (10-Gig) ports
- Qty 4 10-gig Ethernet ports (copper)
- A card slot for an optional faster-than-10-Gig I/O card that can be inserted later)
- All I/O ports use a single switch chip
--- Designed for maximum performance ---

#2: CRS (Bad Boy's little brother)
- Zero 1-Gig ports (no 1-Gig only ports)
- Qty ten SFP+ (10-Gig) ports
- Qty 4 10-gig Ethernet ports (copper)
- A card slot for an optional faster-than-10-Gig I/O card that can be inserted later)
- All I/O ports use a single switch chip
--- Designed to be cost effective ---

North Idaho Tom Jones

I apologize, but the name of the topic about the "high speed router", and not on the switch...

I apologize, but the name of the topic about the "high speed router", and not on the switch...

So , I am assuming that you think/feel that a hardware switch-chip and/or a XEON processor and/or 10-Gig SFP+ and/or 10-Gig copper-ethernet ports and/or a faster than 10-Gig expansion slot would would not make a router run faster for Layer 1, Layer 2 and Layer 3 throughput ?

It´s 01/2019: Any rumors when the new router will be released?

It´s 01/2019: Any rumors when the new router will be released?

Usually big announcements are at the EU MUM.

I would like a CCR 1036 with 4 SFP+

04/2019 is over: Any rumors when the new router will be released?

06/2019 is over: Any rumors when the new router will be released?

No need to ask every three months. Once per three years is more than enough.

Once per three years is more than enough.

LOL