2.5 years someone has made the effort, all memory errors on the zigzehntausenden evaluate Google's servers. The results are - like the hard drive of evaluation - crass. An ECC memory module (ie with error correction) has an average error per year 3751.
This is several orders of magnitude more than any manufacturer. But we first look at the bright side: The temperature plays - like the hard drives - no role in the failures. All the additional cooler for RAM, you can check off so confidently. The errors are not the type of DIMM or the manufacturer dependent.
And that's the bad news: Because the study showed that the error rate depends on the motherboard!
"the study found that error rates were motherboard, DIMM not type or vendor, dependent. This means that some popular mobos have poor hygiene EMI. Route . A memory trace too close to noisy component or shirk on grounding layers and instant error problems "have sloppy about the MoBo manufacturer
Next topic: A company" Danger ", that danger to name, is already a comic idea. particularly if the customer data at risk there. Was not otherwise be expected when making out small soft and Tele Goofy together, but to send just like the contact and calendar data, photos and other personal information of customers to nirvana, but is is a bit rich. One hope for the return of important data, it also does not make up is still unclear how many users are affected. What is clear is that the Affected players are pissed. So much about "Cloud Computing".
first Update (11/10/2009): If this comment does at Engadget, there was one in which incompetence Rudelbums.
- over by MS
- fired competent people
- Cloud as the main data storage
- No backup
- outsourcing the maintenance of central components
- At Conslutants!
second Update (12/10/2009) : Fefe other hand, writes in his blog that the rumor was he had been reported that the SAN would have "spontaneously disintegrated . While theoretically possible, I consider it relatively unlikely.
0 comments:
Post a Comment