ServeTheHome.com

This week saw OCZ accused of fraud, a big storage merger announced and divergent earnings reports from EMC and Western Digital. Read the full story

This week saw a few developments both in the SSD space and in the microserver space.

OCZ Purchases Indilinx

Early this week we learned that OCZ has reached an agreement to purchase the SSD controller designer, Indilinx, for $32 million. From what we know, Indilinx’s next-generation controller will be released well after SandForce’s and the company has been relatively quiet since the Barefoot days. Given OCZ’s push to be a major next-gen drive provider (like Western Digital and Seagate are in the rotating disk world), purchasing a controller company at a low cost does make a lot of sense if the next-gen controller will be marketable. Saving $1-2 in licensing fees on drives can make the $32m investment up relatively quickly. Frankly, the purchase price was somewhat interesting. If Indilinx’s next-gen controller was indeed highly competitive, a Western Digital, Seagate, SanDisk, STEC, or DRAM company would most likely have been willing to pay the relatively low price.

ONFI 3.0 Specifications Published

The Open NAND Flash Interface (ONFI) Working Group published the ONFI 3.0 specifications this week. The new specification doubles the current standard to 400MB/s interface bandwidth which should allow for either the same performance given half the number of controller channels or twice the performance given the same number of controller channels. Probably the other big improvement is the ECC Zero (EZ-NAND) interface. EZ-NAND offloads ECC functions to the NAND die instead of using the controller to perform error checking and correction functions.

Intel Previews their Microserver Strategy

In San Francisco Tuesday Intel previewed their low power server strategy including the very low power portion of the Xeon E3 line and will be bringing Atom lines formally to servers. Intel is under increasing pressure from ARM in the low power CPU space and ARM is making an aggressive push into the low power server market. Also, AMD’s new Brazos platform is pressuring the Atom’s low power x86 dominance so perhaps this was a foreseeable and inevitable move. Unfortunately I did not get to go to the event as I had meetings south of the city. On the plus side, I will be looking at low power Xeons in the near future as well as doing a piece on the AMD Zacate platform for next week.

Discuss these topics on the ServeTheHome.com Forums!

At the request of some users, I am putting together a few bits of storage news from the past week. If the response is positive, I may make this a weekly Friday digest. Read the full story

Shortly after posting my last benchmarks on the dual OCZ Vertex 120GB in Raid 0 setup, I received an e-mail to try the new Intel Rapid Storage Technology (RST) 9.6.0.1014 drivers, which were supposed to enable TRIM in raid. These new drivers were released, taken down, then re-released in March 2010. The big improvement was supposed to be TRIM is enabled for Raid 0, 1, 10, but not enabled for Raid 5. Unfortuneately, this seems not to be working so the best that can be done is the background garbage collection in firmware v1.5 for OCZ’s Indilinx based SSDs. Since I am only working with two 120GB drives at the moment, I decided to give them a shot. Just for reference, the original benchmarks were done with the standard, off the shelf, Windows 7 64-bit RTM drivers.

Read the full story