Slashdot Mirror
BusinessWeek On XORP vs. Cisco
cornfed writes "BusinessWeek is running this article talking about how XORP will take on Cisco's dominance in the router market. The article speculates that XORP could represent the next 'open-source rebellion.' One can only imagine the fallout within the telecommunications industry if an open-source project like this gained traction-- Cisco would not be the only giant to be slain."
Here is more about XORP (the Extensible Open Router Platform), for those that don't know.
AnimeNEXT anime convention
convoluted adj 1: rolled or coiled together; "a convoluted shell" 2: highly involved or intricate; "the Byzantine tax structure"; "convoluted legal language"; "convoluted reasoning"; "intricate needlework"; "an intricate labyrinth of refined phraseology"; "the plot was too involved"; "a knotty problem"; "got his way by labyrinthine maneuvering"; "Oh, what a tangled web we weave"- Sir Walter Scott; "tortuous legal procedures"; "tortuous negotiations lasting for months" [syn: Byzantine, intricate, involved, knotty, labyrinthine, tangled, tortuous]
see xorp's website
It's BSD Licensed, so Yes, MS could take and use it, much like their TCP/IP stack.
Why aren't you encrypting your e-mail?
in other words a project like say... Asterisk?
We already have 5+ HUGE (100k+ DIDs) companies running it and raving about it... what more do you need?
Telcos have alot of dark fibre in the States. Most people assume that's optical fibre...but it's actually moral fibre.
Xorp may be fine for low-end applications, where the cost of hardware is more important than the cost of support and uptime.
But, for any relatively complex network, the tech support offerings of a big player like Cisco becomes very important. And, if they have high performance requirements, the custom hardware in a Cisco or Juniper box is pretty tough to compete with on a general purpose platform.
Even at the low end, it's tough to compete with a Linksys/Cisco box doing basic routing functions. In terms of size, power usage, and noise, a small embedded router box is a much better option than a clunky x86 box running xorp.
Last time I looked, none of the open hardware project had done well. When you think about it, any company that is going to build a high quality 96 port ethernet adapter for a PC with hardware to accelerate security, qos and forwarding is going to end up charging a lot of money for it. Then layer in software customization and support and you look just like any other Cisco competitor.
Working in a Cisco based shop, I can certainly appreciate the need for such a product. I look forward to it.
The only problem I have with the article however is its definition of a switch.
The article states: "Switches determine the most efficient path for everything from streaming videos to e-mails to instant messages." It is not correct, switches are not designed to make such determinations.
The Webopedia says:
(swich) (n.) (1) In networks, a device that filters and forwards packets between LAN segments. Switches operate at the data link layer (layer 2) and sometimes the network layer (layer 3) of the OSI Reference Model and therefore support any packet protocol. LANs that use switches to join segments are called switched LANs or, in the case of Ethernet networks, switched Ethernet LANs.
While a router:
(rowter) (n.) A device that forwards data packets along networks. A router is connected to at least two networks, commonly two LANs or WANs or a LAN and its ISPs network. Routers are located at gateways, the places where two or more networks connect. Routers use headers and forwarding tables to determine the best path for forwarding the packets, and they use protocols such as ICMP to communicate with each other and configure the best route between any two hosts.
Such a misuse of terms, particularly from such a respected national magazine, certainly does not help those of us who have to communicate with the non-technical on a regular basis.
"The way you think it is may not be the way it is at all." St. Oran
So that XORP might be open source, but it has to be commercialized as well.
Of course. That's how open source works: people share the source code, but support, integration, add-ons, and improvements are pay-for-service kinds of deals. In fact, for a lot of open source software, it's the companies offering commercial support that are also paying developers for continued improvements to the open source project.
As for Open Source projects, usually, they are known by "fix-it-yourself".
I'm sure Microsoft's FUD machine likes to present open source projects that way, but in the real world, that is totally wrong. If you want support for your open source projects, you can buy it, and you still end up with lower support costs and lower risk than if you buy closed-source software with support.
check the faq :)
http://www.xorp.org/faq.html#pronounce
I would like to point out that most Cisco Kit is actually running IOS off an embedded PPC CPU.
There was even a project to run Linux on most Cisco routers and switches at one time.
Currently you will see a large majority of Cisco's high end equipment moving to commodity hardware running linux.
Examples of this are the Cisco Content Engine line are embeded linux machines. They are effectively a linux box running a proxy server (isn't squid, but has much of the same functionality).
http://www.mcvax.org/~koen/uClinux-cisco2500/
Only company that I know of right now to actually impliment routing and switching in an ASIC is Nortel. Cisco is all general CPU running IOS which is how you get new features in same old hardware with IOS upgrades.
The program isn't debugged until the last user is dead.
Um...MikroTik also violates the GPL.
:(
I repeatedly requested the kernel sources so I could rebuild for an old cobalt box I had laying around, and they repeatedly refused, saying that they wouldn't support running microtik on anything but their hardware, despite the fact that the kernel sources are protected under GPL.
I reported them, but apparently the only one who can enforce the GPL on the linux kernel is Linus himself, and he isn't interested in enforcing it.
Karma: Chameleon (mostly due to the fact that you come and go).
Today, however, the story is different. In particular, using an inexpensive small form-factory PC (especially one with no moving parts, even a fan), you can have a router for $500 that outperforms a Cisco router costing ten times as much -- and has more features!
MikroTik RouterOS has replaced Cisco as the routing core for my network here in Honduras, where price is much more important than it was back in the States. It handles peer-to-peer throttling, per-IP bandwidth management, MRTG support, nice GUI and command-line interfaces, cool scripting language, and includes all the cool stuff that Cisco does -- policy-based routing, OSPF, various queueing strategies, etc.
Like $1000 for a Cisco branch office router vs. $1000 for a PC with enough memory and processing power and networking cards to run XORP and match the router functionality?
Or perhaps under $30 per port for a fixed Ethernet layer 3 switch at 100Mb?
If you think these machines are "needlessly high cost" then I'm not sure you quite understand network requirements. I'm not saying there aren't places where XORP will be successful, but there are places it can't get to in the forseeable future (at least 3 technology generations). The core of any enterprise network is MUCH more complicated than a single switch and employs much more reliability than can be provided by a PC. Companies still buy IBM mainframes for a reason, and that high end in the routing space will be routers from Cisco, Juniper and similar devices for the forseeable future.
The SMB market? Bring on XORP, they'll be playing with it by the end of the decade.
I worked at Cisco too. I was a contractor, I was paid and treated VERY well. Cisco DOES cut the bottom 10% (saw it happen) but not in every group, and the DO bring H1B's over in droves. They pay the H1Bs well (better than most) and most of them are talented, but the still are paying below standard wages. Cisco will make a penny scream for mercy if it affects production costs. They got guys/gals working there that put in 90 hour weeks in the hopes their stock options will ever get above water. I don't know how many times I heard if the stock hits $XX I'm cashing in and leaving. So, in many cases the rank and file employees are OK with the H1Bs if it saves money, as long as they are not replaced by them!
The above is also absolutely true for Juniper, Foundy, Extreme and Force10 - and, as you point out, Nortel. Switching packets in software hasn't been a standard practice (outside of bugs) in most modern platforms for many years.
CPU's are fast now. Heck, memory speeds are getting very fast. An Opteron might even be able to switch packets between a couple of 10G interfaces at- or near- line rate. Now extend that to a box with 32 10G interfaces in it. You now not only need 320G to the physical interfaces via some number of bus connections, you've also got to be able to move packets in- and out- of memory, maintain routing adjacencies and any other miscellaneous network management tasks ... all in real time. PC's are not built to do this. Outside of real-time extensions Linux/BSD/et al are not built to do this.
Think of it this way - assume an average packet size of 300 bytes. On a one gigabit ethernet interface this represents something on the order of 40,000 packets per second ... in one direction. Multiply this by 10. Now by 32 interfaces. What does an OS and PC platform look like that can malloc() 25.6 million times per *second* above and beyond any other OS processes? Oh - and don't forget the fancy queues, packet re-writing, CRC calculation and such that would necessarily follow each one of these memory operations.
This is obviously an extreme example, but it illustrates the point that everyone in the industry pretty much figured out a bunch of years ago that distributed forwarding via dedicated hardware was the only realistic answer to this problem. This is why the CPU in just about any Cisco platform you'll see in common is less capable than a lot of PDA's out there and also why the architecture of basically all of the major players is moving toward a condition where forwarding and network control are handled by roughly autonomous units.
ASIC= Application Specific Integrated Circuit, and yea, Cisco's stuff is chock full of em.
Router# show mls asic
Cafe version: 2
Centauri version: 1
Perseus version: 0/0
Titan version: 1
Clip from Cisco.com
As technology and features mature, they often move from a software-based implementation to inclusion in hardware. At the core of Cisco's hardware integration is application-specific integrated circuit (ASIC) innovation. Cisco has developed more than one hundred ASICs for the Cisco Catalyst switching family over the past nine years, with each generation including more capabilities. For example, Cisco was the first vendor to integrate Layer 3 switching into hardware with the Cisco Catalyst 5500 NetFlow Feature Card. With the introduction of the Cisco Catalyst 6500 Series in 1999, Cisco included features such as quality of service (QoS) classification and queuing and security access control lists (ACLs), and provided them at data rates of millions of packets per second. These features are available, in hardware, across the Cisco Catalyst switching product line, including the Cisco Catalyst 4500, 3750, and 3560. Advanced hardware integration continues with the Cisco Catalyst 6500 Series Supervisor Engine 720, which integrates MPLS, IPv6, and generic routing encapsulation (GRE). This is the first time a LAN switch has offered this capability at data rates in the hundreds of millions of packets per second.
Never ascribe to malice what can be adequately attributed to ignorance. -Napoleon
You have switches with DSL interfaces, with modem interfaces, with IPsec built in?
What make are they?
Now, assuming that they are in fact violating the GPL, anyone who has copyrighted material in the work can in fact force their hand. So any kernel developer can deal with this.
Looking on their site, they have in fact given lipservice to sending you a CD (they claim it won't contain their propriatary software, but if they are following the letter of the law, they should have to give you the kernel source w/ any modifications they made). The offer is down near the bottom of this page:
http://demo.mt.lv/help/license.html
This appears to be in compliance with 3b of the GPL.
If you report this on the LKML list I'm fairly sure several people would help you pursue it if you can show they are in fact violating the GPL (if they didn't modify the kernel, they aren't). If they are violating the GPL, they sure are being quiet about it. Google turns up very little about it. I've seen several threads on the LKML where people outside of Linus Torvalds pursue GPL violations. Alan Cox being on of them. Any number of people pursued Linksys.
http://lkml.org/lkml/2003/6/7/164
and
http://linux.derkeiler.com/Mailing-Lists/Kernel/20 03-09/7435.html
are examples
Kirby
As usual, the most uninformed post is at +4 informative.
Practically all of Cisco's gear uses custom ASICs. You don't switch at near-wire-speed with a general purpose cpu. You use a switch fabric implemented on an ASIC, and nearly all of Cisco's are done in-house. The PPC runs the CLI, programs registers in the ASICs, and handles low-bandwidth protocol work, e.g. RIP, or 802.1x. Most packets are never seen by IOS.
Layer 2 or 3 switches? Since layer 3 features have crept into the switch fabric, the distinction between "router" and "switch" has blurred. You can buy "switches" that route layer 3 interfaces, and "routers" that have switch blades.
If you're really running a flat network, all at layer two, the obvious answer is that it doesn't scale. In fact, there is momentum in the other direction, to push layer 3 into the switch closet because it avoids common layer 2 problems, and the switch fabric is the same speed whether it is switching packets based on mac or ip.