Slashdot Mirror
Unwinding Cisco's Not-So-Simple Beginnings
saridder writes: "There's a saying that behind every fortune is a crime, and
as we have learned with Apple, Microsoft, and others, Cisco is no
different. The SJ Mercury
has an article outlining and debunking the myth of Cisco's
founding."
Did you read the article at all? They took software written by other Stanford people, loaded it onto PCBs they were having built at Stanford's shops, and selling them for their own profit. The only "opportunity" that was capitalized here was the opportunity to rip off a lot of smart and hard working people.
A point lost in the minds of many who argue for patents is that very few ideas, if any, spring from the mind of one individual. In reality all high-tech ideas leverage the ideas of many who never get attribution. That's just a fact.
Do I argue for patents or against? If you believe what I just stated, which do you think?
Cisco is just one example getting a little more light than usual.
--- -- - -
Give me LIBERTY, or give me a check.
... is universities using students as slave labor.
Universities engage in scientifically uninteresting joint projects with the industry to raise money. They employ students as free labor. The students get units, the lab gets unearmarked funds and the industry gets cheap software.
The head of our lab called it prostitution. Since most students don't have a choice, you might as well call it forced prostitution.
First, all the major network developers saw the need for routers. Xerox PARC had had PUP/XNS routers for a while, based on Alto machines. PDP-11/34 minicomputers running Dave Mills's "fuzzball" code had been routing IP datagrams since 1981 or so. BBN had several routers. I built an IP router myself at Ford Aerospace in 1985, using a VMEbus cage with a Motorola 68000 and some Ethernet boards. I'd previously had a VAX doing routing in its spare time. So how to do it was understood.
I knew about the Stanford routers, but felt that their Multibus card cages weren't solidly enough built for deployment. (Remember, I was at an aerospace company.) Commercial VMEbus stuff was starting to appear, and that seemed the way to go, even though it cost more.
We were trying to get away from multiprotocol routers, which add an extra layer to everything. We were thinking "TCP/IP everywhere", rather than routing SNA, DECnet, XNS, X.25, and TP4 (all of which have been forgotten) over the same wires.
But a mass market for routers seemed a long way off. Ford Aerospace had built some big digital networks for DoD in the past, and they typically had 10 to 100 switching points. Management didn't see a case for a volume product. (Ford Aerospace had been badly burned on some previous products that were too early, like a really nice projection TV in the 1970s).
The major vendors were all fighting TCP/IP in favor of their proprietary network protocols. This was the era of the "PC LAN". Ungermann-Bass, Network Systems, and 3COM all had incompatible PC LANs. IBM had three PC LANs which wouldn't talk to each other.
Cisco was more of a marketing success than a technical one. There was no real obstacle to building a router by 1984 or so. But selling lots of them looked hard.
At first I thought about it and thought you were right, but the more I think about it, the more I have to disagree. There are plenty of knock off routers and switches already (3com, Linksys, Netgear, etc.,) and Cisco is still number one in the industry. Plus I looked at your link and they made almost a billion dollars profit each quarter! That's a lot of PROFIT. The 2 billion loss was for a write off of some equipment the they produced for the now defunct telecom and ISP's that fell through last year.
When you think of routers, don't just think of the cheap toys sold in CompUSA and BestBuy for mom and pop at home. A major corporation, ISP, web company, financial institution, etc., is not going to base their critical network infrastructure on a cheap knock off, nor could a cheap knock off handle a high load. Imagine an ISP running off of a Linksys.
And even of a company made a cheap alternative to compete with a 7000 series router (never mind an optical router), once all the R&D is done to make the big router, create all the software to comply with RFC's, create all the hardware to handle the different networking technologies, create innovative switching technology to handle the high load of packets, the router isn't cheap any more. (Look at Juniper, ArrorwPoint, etc,.) And if it works, Cisco will just buy them, paint the box blue and call the Cisco 54000 or something J.
Routers and switches do more than just route. There are many technologies that just can't be implemented cheaply. When I think of a major network, even enterprise, there is a huge need for a Catalyst 6500, especially when there are a ton of users on that floor. Buying cheap 24 port 3coms won't scale, and can't route between Vlan's. Plus there's a ton of other technologies that a company needs - gig backbones, multilayer switching, STP, layer 4-7 load balancing, high speed backbones, DiffServ and COS QoS, and ton of others. Let's see a cheap toy do that.
--- RFC 1149 Compliant.