Re: Yankee Group declares core routing obsolete (was Re: Anybodyusing GBICs?)
Thus spake Daniel Golding [EMAIL PROTECTED] Hmm. Don't you just love it when folks say things like Layer 3 Switches are better than routers. Its very illuminating as to clue level. I suppose what they were trying to say, is that products that were designed as switches, but are now running routing code, are superior to products that were designed as routers, and are running routing code. Of course, this is demonstrably false. Layer 3 Switch is like Tier 1 ISP - meaningless marketing drivel, divorced from any previous technical meaning. I've always stated that switch is a marketing term meaning fast. Thus a L2 switch is a fast bridge and a L3 switch is a fast router. In this light, the Yankee Group is just now catching on to something we all knew a decade ago -- slow (i.e. software) routers are dead. There's a more interesting level to the discussion if you look at what carriers are interested in for their backbone hardware today; while I'm obviously biased based on my employer, I've seen a lot more emphasis on $20k-per-10GE-port L3 switches than $200k-per-10GE-port core routers in the current economic climate. S Stephen Sprunk God does not play dice. --Albert Einstein CCIE #3723 God is an inveterate gambler, and He throws the K5SSSdice at every possible opportunity. --Stephen Hawking
Re: Yankee Group declares core routing obsolete (was Re: Anybodyusing GBICs?)
Funny I thought a switch was a multiport bridge... uses the MAC headers to flood. ahh makes me long for the days of Kalpana. Scott C. McGrath On Fri, 31 Oct 2003, Stephen Sprunk wrote: Thus spake Daniel Golding [EMAIL PROTECTED] Hmm. Don't you just love it when folks say things like Layer 3 Switches are better than routers. Its very illuminating as to clue level. I suppose what they were trying to say, is that products that were designed as switches, but are now running routing code, are superior to products that were designed as routers, and are running routing code. Of course, this is demonstrably false. Layer 3 Switch is like Tier 1 ISP - meaningless marketing drivel, divorced from any previous technical meaning. I've always stated that switch is a marketing term meaning fast. Thus a L2 switch is a fast bridge and a L3 switch is a fast router. In this light, the Yankee Group is just now catching on to something we all knew a decade ago -- slow (i.e. software) routers are dead. There's a more interesting level to the discussion if you look at what carriers are interested in for their backbone hardware today; while I'm obviously biased based on my employer, I've seen a lot more emphasis on $20k-per-10GE-port L3 switches than $200k-per-10GE-port core routers in the current economic climate. S Stephen Sprunk God does not play dice. --Albert Einstein CCIE #3723 God is an inveterate gambler, and He throws the K5SSSdice at every possible opportunity. --Stephen Hawking
RE: Yankee Group declares core routing obsolete (was Re: Anybodyusing GBICs?)
. Things are getting better, but L3-switches pale in comparison to today's high-end routers on almost all fronts. If you take GigE out of the equation, modern L3 Switches are just as expensive as modern core routers - and routable, mpls-able L3 GE ports are _more_ expensive on switches than routers (see 4xGE OSM vs 4xGE GSR 'tetra' pricing). Media diversity, queuing performance, and FIB density is what really differentiates the two at this point, IMO. [stuff deleted all over the place] Christian, I think you make the point very clearly, if you leave GigE in the equation things change a lot. Without it, none of this stuff walks too far. GigE is being used in all kinds of IX, LAN, and Metro environments that WAN circuits or at best FE used to be used for. This reduces the number of low speed and short-haul interfaces on most core routers immediately. 10GE still isn't a very far reaching technology yet (meaning, I can't seem to find one stable at 26db) and SONET clearly wins in speed range for distance AFAIK. For networks that can engineer or re-engineer to GE or nxGE an L3 switch is going to do very well. Many support hardware rewrite for L2 forwarding, and newer ones are sporting real-router sized FIBs. Even in an IX environment, if you are only talking to peers, you can use an L3 switch with a 20,000 route FIB and know you'll never be defaulted to, and all of your BGP views at least 100 sessions can be aggregated on a little 1U box that costs $4000. You also protect your main router from a lot of nonsense that can be hw-filtered on the little box. If big routers could provide GE ports in higher densities at approximately the same price per port as a switch, the argument would be a dead one. Its expensive to privately (router) peer with 30 GE networks on a vendor J or vendor C router. Its relatively inexpensive to do it using an L3 switch. When talking about routers that need to aggregate lots of FR, ATM, or other WAN traffic -- or generally uplinking at greater than GE speed interfaces, you are probably better off [today] using a traditional router. I don't think anyone uplinking at 10GE speeds doesn't have a fair about of WAN connections. I don't think most people with lots of GE have many big core routers. I think its a self-selecting type of arrangement. Just my opinion, Deepak Jain AiNET
RE: Yankee Group declares core routing obsolete (was Re: Anybodyusing GBICs?)
Things are getting better, but L3-switches pale in comparison to today's high-end routers on almost all fronts. If you take GigE out of the equation, modern L3 Switches are just as expensive as modern core routers - and routable, mpls-able L3 GE ports are _more_ expensive on switches than routers (see 4xGE OSM vs 4xGE GSR 'tetra' pricing). In *my* Cisco GPL, 4GE-SFP-LC is listed at $75,000 while OSM-2+4GE-WAN+ is listed at $44,000. But then I tend to think of the 6500/7600 as a router... Steinar Haug, Nethelp consulting, [EMAIL PROTECTED]
RE: Yankee Group declares core routing obsolete (was Re: Anybodyusing GBICs?)
I would be interested in seeing, say, a 7609-GSR or better yet 7609-T640 bakeoff. I think that would prove 2 things - 1) you get what you pay for, and 2) purpose-built routers are still better at routing heavy loads with diverse media. Sure, the loaded 640 will be more expensive, but it will most definitely knock the power supplies off the 7609 in general performance. Perhaps the SUP-720 will change that - I look forward to seeing it in our lab, where I may be reconvinced... If you check out the PDF at this URL: www.eantc.de/press/pressreleases/sep03/EANTC-Summary-Report-Cisco-GigE-Catal yst6500-Supervisor720.pdf (I am sure its available elsewhere) You might be surprised about the SUP720 vs T640 performance for general routing loads. Obviously if you have a lot of WAN interfaces the 7600/6500 just doesn't have all of them, but this performance analysis seemed reasonably complete. I have not seen a similar one for the T640. Deepak Jain AiNET
