Xoasis Says SUN's All Wet

Xoasis responds to a letter from SUN Cobalt about the relative merits of the two companies' Web servers.

[Response to SUN RaQs the Benchmark from April 30, 2002.]

[May 9, 2002]

Dear Editors:

We saw the rebuttal from Sun and, as Bill Roth said of us, it in fact "contains a number of inaccuracies."

1) Sun's attempt to reproduce the bench marks that our lab produced were not conducted in the way that our tests show. Our original bench mark clearly states that the Sun XTR unit and our KTX1-1 contained 50 virtual sites. We too receive the similar bench marks if running their exam without those 50 virtual sites. I would ask Mr. Roth what good a Web server's performance is, if it is NOT configured with actual content and customers. I would ask the ISP-Planet readership to write Mr. Roth personally and explain to him how reliable one of their appliances is with 50 sites configured. I would be happy to offer both our box and a Sun XTR box to a 3rd party to further rebut Mr. Roth.

2) I will argue Mr. Roth's point in regards to "discernible differentiators" by merely saying that Sun did not create Server Appliances, Cobalt created Server Appliances, and Sun bought Cobalt and that Sun has been slow to support their appliance product line. In addition, Sun has lost several key Cobalt group executives in recent weeks such as Stephen DeWitt who was the President and CEO of Cobalt pre-Sun-ownership.

3) Sun has used several non-brand-name component vendors in the past and still does (MIPS or Intel clone, anyone?), and the assumption that Sun has always and will always ship product with brand names is simply misleading. Sun does not even lay out hardware vendors in their own data sheets. Sun does not ship the same components in a given product line. You can see this for yourself by merely pulling open the drive casings one of your own XTR's and noting the drives are not Seagate. Readers might also be interested in viewing the RaQ4's data sheet which is available here [.pdf] that clearly explains the processor is not a genuine Intel processor. In addition, Mr. Roth is apparently confused on his own price points and has misstated the base RAQ XTR price by $800.

Questions for Mr. Roth:

1) Would you be willing to allow a willing third party to show perform accurate benchmarking results AND stand by the results?

2) Why have you misstated your product specifications and pricing?

Regards,

Matt Godden
Xoasis Networks, Inc.

P.S. Thought I would point out one more additional issue. Mr. Roth, even when running the test on his own terms, still corroborated the fact that the Cobalt XTR has inferior performance to our lowest base model the KTX1-1.

Does Sun really want to fight over how bad its products actually are? Only a little bit bad? I find it all very amusing.

[ed. note: We hope this dispute will be settled in a third party laboratory which we, at this time, are unable to provide.]