70 likes | 82 Views
This report covers the latest updates from the September 2005 liaison meeting, focusing on the SAS roadmap and goals for 6Gb/s SCSI technology. It discusses important areas like large system builds, error handling, and 3Gb SATA compatibility. The STA recommends incorporating color pull tabs on Mini SAS 4x connectors. Detailed goals and considerations for 6Gb/s SCSI technology are outlined, emphasizing bandwidth aggregation and maintaining compatibility with existing 3Gb implementations. Key questions and objectives for SAS 2.0 development are also highlighted.
E N D
STA -T10 Liaison Report September 16, 2005 www.scsita.org
Next plugfest for week of September 26 at UNH NDA on STA website Needed to attend planning sessions starting in Aug. Focused on large system build, error handling/exception testing, and 3Gb SATA compatibility STA recommends the inclusion of color (Blue) pull tabs on the Mini SAS 4x connectors to denote a SAS connector Updated Roadmap STA Update www.scsita.org
SAS Roadmap SAS 12 Gb/s SCSI SAS 6 Gb/s SCSI SAS 3 Gb/s SCSI Ultra 320 SCSI 2013 2014 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 Note: Beginning of bars denote first plugfest utilizing the technology www.scsita.org
Preserve 3G installments/infrastructure Double transfer rate while improving cost/performance Reduce the number of host/external storage connections/Gb Be compatible with 6Gb SATA within the constraints of the 6Gb SATA usage models (ie. short backplanes, short cables, etc.) Longer backplanes may require additional components or higher negotiated signaling levels Leverage other PHY standards work (EQ, BER, etc) where applicable For example: OIF CEI-6G, etc. Maintain 1.5Gb & 3Gb SATA/SAS Compatibility Same usage models apply for 6Gb as they did for 3Gb Including backplane and cable distances, applications Additional cost burden for rack-to-rack cabling solutions (10m) acceptable, but must not add cost to current usage models Equalization schemes should not appreciably burden cost & power budget, especially for disk drives Critical mass components ready for plugfest testing in mid-2007 Preliminary Goals for 6Gb www.scsita.org
Bandwidth Aggregation If required, no need to optimize for 1.5Gb bandwidth Need diminishes with time Complexity is a risk to meet the market window Disk drive requirements vs. host/controller connections Preliminary Goals for 6Gb www.scsita.org
6Gb Considerations meeting informative-will do it again in Austin Need additional quantitative information on alternatives What are the relative power and die size implications of the different EQ schemes What are the detailed technical impacts (protocol and relative costs) for the different bandwidth aggregation schemes STA believes the same usage models apply for 6Gb as they do for 3Gb Backplane distances, system applications, etc. Are there any technical considerations that conflict with this model? Does the change in the roadmap timeframe change the thinking on the need for bandwidth aggregation? The longer it take 6Gb to enter the market, the less interesting aggregation becomes What the timing difference for a specification with and without bandwidth aggregation? Is there any work that can be done by UNH to help ensure backward compatibility with existing 3Gb implementations or BER? SAS 2.0 Questions www.scsita.org
Circulate objectives within STA Request feedback within 2 weeks Republish objectives with a request for presentations Next Steps www.scsita.org