whats the timetable for wimax on routerboard?

what's wimax ?

seriously though, is there really any future in it ? all I see is press releases and marketing speech, no real consumer products. they say, 4G and LTE is the way to go

ah interesting, when there are performing mini pci cards, then you will support it?
according 'press releases' its being deployed in almost every country

we make affordable yet powerful products. right now, wimax is neither.

It would be interesting to use this 3,5/3,6Ghz frequencies for 11n.
No huge bandwith but good separation from 2,4 and 5GHz.
It's completely unused in our area.

Here you have cards from 450 MHz to 6 GHz.

http://www.xagyl.com/store/home.php?cat=251

3.650-3.675 GHz

http://www.xagyl.com/store/product.php? ... 251&page=1

It's probably only 54 Mbps.

Normis I love you guys. The product is rock solid and you are quick to respond to bugs.

Here is America we Mikrotik users are getting hammered by Motorola Canopy gear. We need something to contend with their product. Contention-based protocol is the future and I do not want to give up our customer base to the cellular carriers. We need proper support for the 3.65 GHz card to fit our Wimax licenses. I have had a license for almost 6 months waiting to see if you guys will develop something. I am about to jump to Canopy, and believe me I do not want to. Our entire network, routers, customers, etc are routerboard based. I and your users want to stay with you. Make the Wimax 802.16 part of a level 7 license or something. I will pay extra and so with others. Because once I drop $20,000 to outfit a tower with Canopy 3.65 gear, I will not look back.

Please consider it so we can remain friends.

Different parts of the planet are suitable for different technologies, mainly due to terrain.

You could drop $20,000 on one AP site where I am (southern Spain) and get the coverage of 1 MT AP.

I dropped about $10,000 on lots of APs in this area and made really good coverage using Mikrotiks.

My point is that WiMAX doesn't really exist as a thing you can buy on the cheap.
You can buy Canopy, Tsunami etc. at a very high price.

WiMAX isn't where Mikrotik is placed in the market.

Basically there's no point trying to push them into a place where they can't exist, because you would not pay even $5k on a Mikrotik RB-WiMAX board and one radio card, or the extra $2.5k for the RB-WiMi second radio, or $1.5k for the RB-WiGPS GPS time synch module.

Ubiquiti are trying it with their AirMax nonsense, and not getting very far.

very intresting...
if there are no other reasons why this stuff is so expensive then we should see it coming into mikrotik range in a year or so..i read somewhere that there is a reference design for a wimax card for under 100$
i am wondering if we just gonna need drivers and OS support for wimax or would a RB433WM52Hn board be necessary?
looking back i remember mini pci 11b cards being the 400$ range at one point

WiMAX isn't where Mikrotik is placed in the market.

but only because it is so expensive. ill say its a good time to play with wimax and push it out as soon as it is a affordable yet powerful product.

You definitely have a point there.

I really doubt that we will se a Mikrotik Wimax Base Station.

In my country, we have at least one Wimax base station in every county. I've never heard any news about Wimax since the deployment of 4G.

I think Wimax is already obsolete, 4G offers 100 Mbit/s to mobile users, and 1 Gbit/s to stationary users.

Pictures of Wimax base Stations.

I think Wimax is already obsolete, 4G offers 100 Mbit/s to mobile users, and 1 Gbit/s to stationary users.

Where can I sign up? I wasn't aware any operators were offering those types of speeds over LTE.

I saw in the news that Elisa and Nokia Siemens Networks started a 4G LTE network with speed up to 100 Mbps in Finland.

http://lteworld.org/news/elisa%E2%80%99 ... ps-finland

In US you can get speed between 40 Mbps and 50 Mbps with Verizon Wireless.

http://www.fiercewireless.com/story/ver ... 2010-03-08

The Swedish company Ericsson set a new mobile data transfer record at the Mobile World Congress in Barcelona with a demonstration of its 1Gbps LTE/4G technology.

http://www.dvhardware.net/article41081.html

The current WiMAX revision provides up to 40 Mbit/s, with the IEEE 802.16m update expected offer up to 1 Gbit/s fixed speeds.

Gigabit ethernet is great, unless everybody has it, then you got to have multi-Terrabit backhaul/distribution.

I don't think we got that yet ...

I saw in the news that Elisa and Nokia Siemens Networks started a 4G LTE network with speed up to 100 Mbps in Finland.

http://lteworld.org/news/elisa%E2%80%99 ... ps-finland

"For pre-commercial use." Not available today. 100mbps to mobile users…is that aggregate from the BTS, or speeds to a single user? If its speeds to a single user I find it very interesting because…

In US you can get speed between 40 Mbps and 50 Mbps with Verizon Wireless.
http://www.fiercewireless.com/story/ver ... 2010-03-08

Those aren't real-world speeds…

"However, Verizon's average LTE network speeds for real-world use did not change from numbers it provided late last year. The company maintains that the network will be able to produce average downlink speeds of 5-12 Mbps and average upload speeds of 2-5 Mbps."

The service isn't even available. The article clearly states these tests are being done before commercial deployment later this year.

The Swedish company Ericsson set a new mobile data transfer record at the Mobile World Congress in Barcelona with a demonstration of its 1Gbps LTE/4G technology.

http://www.dvhardware.net/article41081.html.

And they had to use a total of 80mhz of spectrum to achieve those speeds.

"The demonstration uses four carriers of 20MHz each, totally 80MHz and 4x4 MIMO, data is sent over the air-interface on four independent bit-streams."

According to the following link 80mhz is 88% percent of both AT&T and Verizon's total available spectrum in the US, and its about 48% more spectrum than T-Mobile or Sprint even own.

http://www.sidecutreports.com/2010/03/2 ... -the-most/

You're never going to see those speeds to a single fixed consumer site. It doesn't make business sense to wipe out that much spectrum for a single user, and the operators realistically don't have enough available spectrum to provide that service.

The current WiMAX revision provides up to 40 Mbit/s, with the IEEE 802.16m update expected offer up to 1 Gbit/s fixed speeds.

I think its important to note that these are likely aggregate speeds. I know that's the case with 802.16d as I run such a network, and can verify those aggregation speeds.

I'm not particularly for or against WiMAX or LTE. I think its important to accurately state the performance of each network, and quit trying to downplay WiMAX by making erroneous statements such as LTE will be able to deliver 100mbps speeds to mobile users. Both technologies are evolving at a rapid pace and WiMAX still has a place in this world.

That said, WiMAX equipment manufacturers and providers have already caught on that LTE will likely be the dominant 4G technology due to the backing by the larger cell carries, and as such are already positioning their products to allow or an easy upgrade to TD-LTE.

http://www.maravedis-bwa.com/Issues/5.29/Readmore3.html

Do the maths.

Assume just 100 USA Users all with 1Gb/s Download (upload 128kbit

100Gb/s between USA and, say, Indonesia ?

Multiply that by 10 million. 10Pb/s link required, as a *part* of a network - not a dedicated link.

Come on. Dunt work Today, or even tomorrow.

Better that they download less intensive porn really, 'cos current techology cannot support it.

In europe we have those speeds by the way. 100-500Mbit per home user. It's not wireless, but if the infrastructure allows it, what's the problem to convert the endpoints to wireless then?

"For pre-commercial use." Not available today. 100mbps to mobile users…is that aggregate from the BTS, or speeds to a single user? If its speeds to a single user I find it very interesting because…

Not available today, do you know for a fact or are you just speculating? I can tell you that my operator lauched LTE in december 2009, and it is already available in 25 cities where i live.

Those aren't real-world speeds…

"However, Verizon's average LTE network speeds for real-world use did not change from numbers it provided late last year. The company maintains that the network will be able to produce average downlink speeds of 5-12 Mbps and average upload speeds of 2-5 Mbps."

The service isn't even available. The article clearly states these tests are being done before commercial deployment later this year.

Tell me where you get real world speeds. First, and most importantly, real-world speeds are notoriously difficult to pin down, and vary wildly based on factors including distance from cell site, general load on the network, backhaul and a host of other factors. Further, it's worth pointing out that the speed numbers for each carrier in the US have plenty of caveats. However, Verizon Wireless CTO Tony Melone said the carrier is "going to be aggressive in our plan to get to critical mass" in LTE, so the service will be available soon in the US.

And they had to use a total of 80mhz of spectrum to achieve those speeds.

"The demonstration uses four carriers of 20MHz each, totally 80MHz and 4x4 MIMO, data is sent over the air-interface on four independent bit-streams."

According to the following link 80mhz is 88% percent of both AT&T and Verizon's total available spectrum in the US, and its about 48% more spectrum than T-Mobile or Sprint even own.

http://www.sidecutreports.com/2010/03/2 ... -the-most/

You're never going to see those speeds to a single fixed consumer site. It doesn't make business sense to wipe out that much spectrum for a single user, and the operators realistically don't have enough available spectrum to provide that service.

There is no need to give those speeds to end consumers, most of them probably have reliable wired lines with speed up to 1 Gbit/s or more.

I think its important to note that these are likely aggregate speeds. I know that's the case with 802.16d as I run such a network, and can verify those aggregation speeds.

I'm not particularly for or against WiMAX or LTE. I think its important to accurately state the performance of each network, and quit trying to downplay WiMAX by making erroneous statements such as LTE will be able to deliver 100mbps speeds to mobile users. Both technologies are evolving at a rapid pace and WiMAX still has a place in this world.

That said, WiMAX equipment manufacturers and providers have already caught on that LTE will likely be the dominant 4G technology due to the backing by the larger cell carries, and as such are already positioning their products to allow or an easy upgrade to TD-LTE.

http://www.maravedis-bwa.com/Issues/5.29/Readmore3.html

Who are downplaying WiMAX, or give erroneus statements? WiMAX and LTE are two systems developing along the same lines but optimized to work somewhat differently. WiMAX is primarily aimed at Greenfield (new) fixed to mobile deployments while LTE is mostly aimed at incumbent (existing) deployments that must work with existing networks and business practices”.

The next versions of both, 802.16m WiMAXm and LTE-Advanced, are being designed to meet the same guidelines for IMT-Advanced which calls for an adaptive framework that can be used from local area fixed networks to large scale mobile networks and to use multiple carriers across multiple bands of spectrum. No sense getting into the details but it is important to view both WiMAX and LTE headed to become the ‘Swiss Army knives’ of wireless. Should they merge? They are already on the path of converging at many levels and will eventually be practically merged.

Not available today, do you know for a fact or are you just speculating? I can tell you that my operator lauched LTE in december 2009, and it is already available in 25 cities where i live.

The article you referenced clearly stated the service was not available in that area. My comment was made in response to that. I'm glad LTE is available somewhere in the world…

Tell me where you get real world speeds.

You stated that it were possible in the US to get LTE service in the 40-50mbps speed range; I quoted you saying that earlier. The article very clearly states otherwise. You cannot expect me to believe that a single mobile user with a laptop card can achieve 40-50mbps wirelessly with today's technology…

However, Verizon Wireless CTO Tony Melone said the carrier is "going to be aggressive in our plan to get to critical mass" in LTE, so the service will be available soon in the US.

This statement has nothing to do with what they're going to able to really deliver.

There is no need to give those speeds to end consumers, most of them probably have reliable wired lines with speed up to 1 Gbit/s or more.

Then why would you previously state it were possible to offer 1gbps service to stationary users when you knew: 1. its not technically feasible with current spectrum limitations, and 2. people have ethernet or fiber which sufficiently deliver those speeds?

Who are downplaying WiMAX, or give erroneus statements?

You said "I think Wimax is already obsolete." Is that not erroneous enough for you?

Comparing your last two posts and this most recent, I feel like I'm talking to a different person. The last section of this most recent post is a bit more level-headed and honest with the state of WiMAX and how it is technically progressing. Refreshing.

You stated that it were possible in the US to get LTE service in the 40-50mbps speed range; I quoted you saying that earlier. The article very clearly states otherwise. You cannot expect me to believe that a single mobile user with a laptop card can achieve 40-50mbps wirelessly with today's technology…

I have had LTE with up to 40 Mbp/s throughput since the end of last year. Not everywhere, but generally faster than Turbo 3G (HSDPA). In the same article, they wrote that Verison plans launch to 25-30 Mbp/s LTE this year.

Then why would you previously state it were possible to offer 1gbps service to stationary users when you knew: 1. its not technically feasible with current spectrum limitations.

I didn't say that one single user user will get speed of 1Gbp/s, but technically it is possible. However, if I post a link to an article, it's not my statement, but the author's statement.

You said "I think Wimax is already obsolete." Is that not erroneous enough for you?

Prove to me that this statement is not erroneous. Give me a list of operators with WiMAX. I don't think there is many of them, not even in the US. WiMAX hasn't even been deployed properly and Clearwire could ditch it for LTE standard. According to a Clearwire executive, a change in the terms of an agreement Clearwire has with Intel will allow "either party to exit the agreement - which had until now forced Clearwire to use WiMAX through Nov. 28, 2011 - with just 30 days notice".

I think I read somewhere that Clearwire will run WiMAX until LTE is rolled out, after that they will decide the future of WiMAX.

In Europe, WiMAX players have found themselves in retreat, a state of play which is likely to deepen in coming months. In the UK, Freedom4 recently sold its WiMAX licences to a rival ISP, having suffered losses of £800,000 annually. Fundamentally, the company could not attract enough customers per base station to make ends meet.

In the Netherlands, Worldmax’s Aerea service in Amsterdam is to close in August. The company cited its inability to expand its service due to interference in the 3.5GHz band with a ministry of defence ground station, but there are a number of more pressing factors, including the company’s failure to secure a 2.6GHz licence, the lack of WiMAX devices, the economic downturn which has obliged consumers to tone down their spending, the network’s lack of roaming and poor indoor coverage.

Across the board, WiMAX operators must come to terms with the availability of competing mobile services (HSPA+ and LTE). This will be especially true during 2011 when a number of commercial services are expected to be launched. With every passing month that LTE shows itself as a tried and tested technology, investors will be tempted to swim with that stream as surely as will consumers.

Map of mobile WiMAX service around the world

http://www.muniwireless.com/maps/

According to a Clearwire executive, a change in the terms of an agreement Clearwire has with Intel will allow "either party to exit the agreement - which had until now forced Clearwire to use WiMAX through Nov. 28, 2011 - with just 30 days notice".

They're going to be stuck with WiMAX for a while. They are majority backed by Sprint, who wanted to be first and went with WiMAX. Even though their CEO at CTIA 2010 said he thinks LTE will be the bigger player and that Spring may eventually migrate, they are - for now, and for the medium range future - stuck with their investment in WiMAX. They'll rather partner more with Sprint than less, and maintaining both technologies on all towers simply doesn't make sense.

LTE is between 4mbit/s - 12mbit/s speeds.

Wimax is the same too.

None of these technologies can be called as 4G because the specifications for 4G are 400mbit/s (not sure ) for mobile and 1gbit/s for fixed clients.

These 2 technologies cant do what 4G specs says.

Also either with AirMAX or nstreme (802.11n) or whatever the upcoming TDMA that mikrotik will release soon, you will be able to supply at least 4mbit/s per client even on busiest APs. So if you supply higher fixed speeds by these technologies you are in the game.

My ISP started to upgrade LTE to 100 Mbit/s three months ago. Yesterday, I went to the city for a test, my speed tops out at 80 Mbit/s.

Did Anyone try use XR7 miniPCI cardon RouterBoard? Does this card work on routerboard? I want to 700MHz spectrum on Routerboard.

What i really really cannot understand is why anybody using, well, cheaper gear would imgaine that they can compete on the same level with people using much more expensive gear.

Currently the 'fasion' is for super-duper-fast (100Mb/s+) internet everywhere.

It is a Lie. Those speeds aren't available anywhere apart from the big hubs.

IMHO it is a sales push that will ultimately fail, and quite quickly.
The ROI simply doesn't exist for a start.

Somebody tell me where to buy the kit that will allow 1 million London users, each at 100Mb/s (download only) simultaneously watching a 3D live video feed from a server based in Texas.

Or at the very least explain how that would work, at a profit.

when you say "anywhere", mention the country

we have 200-500mbit FTTH in Latvia, and those are real speeds, international traffic. Of course depends on the connection on the other side, but downloads from the US often reach 100Mbit.

What i really really cannot understand is why anybody using, well, cheaper gear would imgaine that they can compete on the same level with people using much more expensive gear.

Currently the 'fasion' is for super-duper-fast (100Mb/s+) internet everywhere.

It is a Lie. Those speeds aren't available anywhere apart from the big hubs.

IMHO it is a sales push that will ultimately fail, and quite quickly.
The ROI simply doesn't exist for a start.

Somebody tell me where to buy the kit that will allow 1 million London users, each at 100Mb/s (download only) simultaneously watching a 3D live video feed from a server based in Texas.

Or at the very least explain how that would work, at a profit.

Troll: Anything that 1 million london users will want to see at the same time is delt with via CDN's. You really think that the youtube video or apple trailer you watch is coming from CA? 100mbit+ last mile is coming because in a few countries now its being pushed by the goverment. Of course your not going to have Petabyte transatlantic links but you dont need it, the content you users will want on that level will be handled by CDN's. Streaming video, Tv, Music etc is easily done by multicast. On demand content that's popular can be pre-cached to STB's during network useage lulls.

Just look at Japan and Korea and the likes and you'll see 100mbit+ connections being offered that get that speed, Your way of thinking about progress like last mile speed boosts and v6 will make you a relic and someone will come along who will be faster than you for less cost

when you say "anywhere", mention the country

we have 200-500mbit FTTH in Latvia, and those are real speeds, international traffic. Of course depends on the connection on the other side, but downloads from the US often reach 100Mbit.

While this is not a FTTH forum, I thought i would ask a question - because I am changing from a wireless network for a FTTH network

What do you get when you perform up/down bandwidth speed tests to outside internet speed test locations?

Here in North Idaho (good ol USA), I am considering a 3,500 client install of triple-play FTTH termination boxes which has phone, Internet and IP-TV set top boxes.
I am considering adding a (Mikrotik ???) AP inside the FTTH termination box. The wireless router could then NAT/DHCP the customer network and the customer would have the ability to connect up to the Ethernet or connect to a secure SSID (bridged ethernet and secure SSID wireless) we configure for each customer.

Here is the potential neat part - The primary SSID is the secure customer network and I am considering using vlans on multi-SSIDs outside of the customer bandwidth they are paying for. These additional SSIDs (not part of the customer network) could be used for a multi-city wireless VoIP network and would have the most APs in any hot-spot network. It would be possible for wireless network devices to roam the streets in cities without ever loosing their wireless connection because they could roam from one of my 3,500 APs to another AP (just a house away). Kinda like a wireless cell phone where you can be on a call at home and drive to work and keep your call the entire trip. So another question I have for you is have you ever heard of anybody doing something like this in a FTTH network?

Tom Jones - A WISP in North Idaho

It would be possible for wireless network devices to roam the streets in cities without ever loosing their wireless connection because they could roam from one of my 3,500 APs to another AP (just a house away). Kinda like a wireless cell phone where you can be on a call at home and drive to work and keep your call the entire trip. So another question I have for you is have you ever heard of anybody doing something like this in a FTTH network?

Unless all those APs are bridged to the same network or steered through a controller for roaming through proprietary extensions (expensive, not possible with MT) there'd be interruptions for the clients when they change APs and end up on a new gateway (get a new DHCP lease, all existing connections drop, services that also use the WAN IP as seen by the server as part of the session log the session out). That may work for casual browsing but would critically interrupt streaming services for video and audio, which many people use nowadays on mobile devices. And one thing people/customers appear to like less than having no service at all is a service they experience as flawed.
Just playing devil's advocate. It would be a fun project.

I remember reading about an ISP (or just some project?) in Europe where wireless edge routers given to customer also broadcast a secondary SSID that the user can opt to share out. It becomes usable immediately for all other customers that also opted to share their service out, giving you the ability to roam across thousands of APs by merit of participating in the roaming network.

Re "Somebody tell me where to buy the kit that will allow 1 million London users, each at 100Mb/s (download only) simultaneously watching a 3D live video feed from a server based in Texas.
Or at the very least explain how that would work, at a profit."


This is almost my plan - we are installing FTTH. We will be using a G-PON network with no more than 16 homes per PON pool. That comes out to 16 homes per PON pool being able to sustain 100 meg at the same time or one home sustaining 1-gig. My internet feed will only be 1-gig to run the entire network. Sooo my plan will include rate limiting queues which allow bursing to 1-gig but dropping down as the customer sustains bandwidth to a final lowest bandwidth somewhere between 10 to 100 meg.

- If queue counters are empty - then allow 1-gig/sec down for 1-second
- then allow 500 meg/down for 5 seconds
- then allow 250 meg/down for 60 seconds
- then allow 100 meg/down for 2 minutes
-... this dropping of customer bandwidth continues until we reach the dedicated download speed the customer is actually paying for (it may be between 3mb down to 100mb down for the final sustain rate).

I could not afford to bring in multi 10-gig networks for my FTTH project - but I can use some good rate limiting queues to really stretch a 1-gig internet feed to alot of customers.

note - on my WISP network I am doing something like this now - a low-end client with a 200k account can actually do a speed-test and see 20-meg down for a few seconds. but try and sustain it and their bandwidth drops way off. I have some 2.4 APs with over 100 connected clients - and this works out ok - unless they are trying to watch a movie. Short YouTube videos really pop in pretty quick on a 200k wireless customer.

re "...Unless all those APs are bridged ..."Yup....

I am thinking about putting all of the secondary SSIDs on a common vlan network wide.
SSID #1 - secure customer network (local customer network only)
secondary SSIDs - all secondary SSIDs are bridged to a common vlan network wide. Instead of roaming between just a couple of APs, I would be able to roam between 3,500 APs.

The gateway and dhcp server for all 3,500 APs (secondary SSIDs) would actually be in our NOC. A clinet dhcp assigned IP address would be network wide. There is only one dhcp server and one gateway and only one IP network in the entire secondary SSID network.

Re: "I remember reading about an ISP (or just some project?) in Europe where wireless edge routers given to customer also broadcast a secondary SSID that the user can opt to share out. It becomes usable immediately for all other customers that also opted to share their service out, giving you the ability to roam across thousands of APs by merit of participating in the roaming network."

The secondary SSIDs would not be part of the speed or bandwidth a customer is paying for. The queues for secondary SSID would be outside and seporate from the queues the customer is actually paying for. Thus if I have a 10 to 100 meg customer - I still have over 900 Meg to operate the secondary SSIDs per FTTH device.

Even if a clinet is saturating their bandwidt - the bandwidth running the secondary SSIDs is seporate and a diferent network. A client could even cancel all service from us - but because the FTTH box is still mounted there - the secondary SSIDs would still operate.

1-gig

How is that possible on TCP/IP ?
More specifically, how is that possible to actually deliver that with current CPU technology - never mind the radios.

With 16 simultaneous users over 1 second, that's 16 Giga bits in a second = 2 Giga Bytes in 1 second.

If the CPU of the first router is at say 3Ghz, and assuming an instruction cycle of say 4 cycles (so a RISC CPU architecture, not x86), and a TCP/IP stack overhead of just 20 instructions (!) then the CPU could handle a max of 37,500,000 bytes in 1 second, so a maximum possible thruput of 2,343,750 bytes per second per user = 18,750,000 bits per second per customer if there are 16 simultaneuous customers.

Or 19 Mb/s thereabouts, rather than 1 Gigabit per second, as a hard limit given what we got.

1-gig

How is that possible on TCP/IP ?
More specifically, how is that possible to actually deliver that with current CPU technology - never mind the radios.

With 16 simultaneous users over 1 second, that's 16 Giga bits in a second = 2 Giga Bytes in 1 second.

If the CPU of the first router is at say 3Ghz, and assuming an instruction cycle of say 4 cycles (so a RISC CPU architecture, not x86), and a TCP/IP stack overhead of just 20 instructions (!) then the CPU could handle a max of 37,500,000 bytes in 1 second, so a maximum possible thruput of 2,343,750 bytes per second per user = 18,750,000 bits per second per customer if there are 16 simultaneuous customers.

Or 19 Mb/s thereabouts, rather than 1 Gigabit per second, as a hard limit given what we got.

--> Re 1-gig - We will have a 1-gig fiber connection to the internet from our Cisco 6509 router. The fiber connection will be able to support an upgrade to a 10-gig fiber link when needed. The Cisco routers we have support it now.

As far as distribution of that bandwidth to the customers - currently we have about 100 APs - However when we have our FTTH installed the fiber to the home has a peak throughput of almost 2-gig per Pon network. We are planning on ratelimiting each customer to a max burst anywhere from 100Meg to 1-Gig then allow bursting rules in the customer CPE device auto-throttle the sustained bandwidth down to what the customer is actually paying for. (note - customers will have a 10/100/1-gig hand off at their homes and a second private wireless SSID at the same time - all part of the same customer IP network).

Not customers can not directly sustain greater than our 1-gig bandwidth - but with customer bandwidth bursting any customer can burst to really high speeds for a few seconds then auto-throttle back down to their purchased speed.

Also - as an option - a customer can set their computer to use our caching proxy server (squid) - it will have several 10-gig network interfaces and be able to cache up to about 2-terra-bytes of popular web content. This greatly reduces our load on the Internet because the 2nd and additional customers going to the same page will auto grab the content from our squid server instead. I have been using squid off and on for over 10 years now. It works OK for almost everything (almost...).

The Cisco routers we have support it now

Cisco still have to use existing technology. They do not have Magic Pixies inside their equipment.

How is 1 Gig or 10 GigaBit per second actually made possible in a router ?

Skill and passing it to silicon designed to do one thing really really really fast.

Seriously, This is nothing new. We've had 10gig routers for a number of years

The Cisco routers we have support it now

Cisco still have to use existing technology. They do not have Magic Pixies inside their equipment.

How is 1 Gig or 10 GigaBit per second actually made possible in a router ?

You have to avoid the normal CPU. Look at graphic Cards. You have to
Design Chips which do Special Operations in parallel and with less cycles.

So, no-one really knows, but they do know that 10Gibabgit Routers exist.

Hmmm.

I wonder if my throughputs will increase if i write 1Pb/s on the top of my RB750.

I'll try it and see if the Cisco Magic Pixies decide to come and live inside my router.

No Magic Pixies arrived yet.

Googling was better : http://www.bbn.com/resources/pdf/50G.pdf

We just told you, custom designed silicon that does nothing but route traffic really well and fast

We just told you ...

I was really asking if anyone knew how it was done.
Now if it were a <0.5um process using parellelised data channels and parrallel cpus, then that'd be an answer.

If it were FPGAs (like Juniper) re-flashing/configuring the logic arrays to make a super fast router, then that'd be another.

Custom silicon = Magic Pixie.

We just told you ...

I was really asking if anyone knew how it was done.
Now if it were a <0.5um process using parellelised data channels and parrallel cpus, then that'd be an answer.

If it were FPGAs (like Juniper) re-flashing/configuring the logic arrays to make a super fast router, then that'd be another.

Custom silicon = Magic Pixie.

Yes Sir!

We just told you ...

I was really asking if anyone knew how it was done.
Now if it were a <0.5um process using parellelised data channels and parrallel cpus, then that'd be an answer.

If it were FPGAs (like Juniper) re-flashing/configuring the logic arrays to make a super fast router, then that'd be another.

Custom silicon = Magic Pixie.

It's Cisco's version of 11 herbs and spices (KFC reference), It's what separates the the Cisco's from the Mikrotik's

The sticker on my RB750 worked !

A magic pixie moved in, and i got 500Gb/s thru it for 12 hours.

I was so excited that i forgot to feed the Pixie, and it left.