On 08/30/99 16:41:59 you wrote:
On Mon, 30 Aug 1999 sirbruce@ix.netcom.com wrote:
A cautionary note - much of the supposedly "in-spec" memory Kingston or other vendors sell is not capable of handling the strict timing and loading requirements of a Netapp filer.
This sounds like FUD. In fact it sounds like something Kingston would say about other memory vendors. Do you have data to substantiate this?
I'm sure Netapp might provide the data under NDA if you asked. I was speaking personally, however. I have *personally* seen tested Kingston memory come in and fail during internal testing on a filer. I have *personally* seen Netapp test filers they build for customers and have the memory in them fail and have to be replaced and re-tested before shipping to the customer. I have *personally* seen filers experience crashes and spontaneous reboots due to having the right memory in the right slots. I have *personally* seen the poor timing traces on oscilliscopes as EE's examines the problem.
I'm not saying don't buy memory from NetApp. We buy it there and will continue to buy it there because we want to maintain our service agreements. However, since you're making claims that memory sold by NetApp is superior please show me the data from an independent source.
How could an independent source exist? You ask the impossible. But if you could get the data from all customers and somehow verify the source of the memory in them, I think you'd see a difference. It just stands to reason. Kingston sells memory to Netapp that has a certain known possibility of failure. Kingston sells that same memory to you, the consumer. Netapp does additional testing and gets some memory to fail before selling it to you, the consumer. Statistically the memory you get from Netapp has to have a lower failure rate. If the testing didn't do anything, why would Netapp bother? The same goes for the disk drives.
Jeff Sloan has said they've now certified certain "direct from vendor" parts to be as good as Netapp supplied memory. My guess is either they have stepped down their internal testing, or the difference in failure rate has become too minimal to matter.
Bruce