nahanah的部落格

Question: LSI, Dell and Seagate not long demonstrated a 6G SAS make friends for the SCSI Trade Association. Was this a of import milestone?

So: This was the most basic practical protest rally of a 6Gbps SAS system featuring both entity of the I/O alley - server, HBA, SAS expander and disk drive. It was a validation of the growth hard work of all 3 companies maddening to convey the most big-chested retention answer to the labor industry. It paves the way for 6G SAS in late 2009.

Question: What makes 6G SAS desirable to 8G Fibre Channel or 10 GbE?

So: SAS was not designed to regenerate any Fibre Channel or iSCSI. It was to begin with mechanized as the next-generation SCSI, exchange comparable SCSI in DAS solutions. But the largest problem of some GbE and Fibre Channel is charge. SAS is a cheaper interface to implement, and faster too. A banner surface linkage naturally holds cardinal SAS biological layers. Each SAS PHY present is 3Gbps, which gives you 24Gbps. Upping that to 6G delivers a 24Gbps bond.

We are sighted go in SAS on a grownup interface as a replacement for Fibre Channel in lower-price bands. SAS offers in good health show at belittle reimbursement than remaining technologies, but it's minor by telegram distances. 3G SAS superior out at astir 8 meters. With 6G, we're superficial at 10-meter wiring. But the meaning here is that we're multiplication the replacement charge and static deed long distances. With collateral SCSI, the queue lengths got shorter as the information measure went up.

Question: So we could be seeing an all new SAN medication then?

So: We're seeing curiosity in 6G SAS as a switched topology, in blade centers particularly. We're superficial at a controller wares for 3G, and we're heart-rending forward on a wares for 6G. There's too new go in SAS as a adult surface for superficial storage, for the most part as a SAS-to-SAS surface for outdoor RAID boxes. We'll unmoving have FC-to-FC and iSCSI-to-SAS, but the go is there for SAS-to-SAS.

The information is, 6G SAS genuinely doesn't sort cognisance until the total substructure is
G. What's the prickle of having 6G if it's going to chokepoint as in a bit as it hits the 3G infrastructure? With 6G, you get greater steadfastness in the interconnect - a more regular intermingle at a quicker charge. You also get new features similar outcome feedback, leveling and standardized partition that eliminates the sympathy issues SAS had once each merchant implemented sectionalization in their own inventive way.