NetApp - Standards Watch

Ethernet Standards, Plugfests and White Papers

Plowing through the disaster area I call my inbox, I came across the latest newsletter from the Ethernet Alliance. Usually I simply delete these things on the grounds that they are of marginal interest and distract me from doing real work.  However, the EA seems to have gotten newsletters right: not too frequent, with good summary information, and no fluff.  You can find the latest one here.

The news items that particularly piqued my interest were as follows…

Standards

The summary from the IEEE March Plenary meeting showed continued rapid progress on IEEE P802.3ba (the 40G and 100G Ethernet standard) and IEEE P802.3az (Energy-efficient Ethernet). The former has now moved from the initial spec development phase into the working group ballot phase, and looks well on track for ratification in June 2010.  The latter (Energy-efficient Ethernet taskforce) plans to request going to Working Group ballot in July.

Plugfests

The EA’s Data Center Ethernet Subcommittee has scheduled a plugfest focused on Data Center Bridging to be held on May 4-8 at the University of New Hampshire Interoperability Lab in Durham, NH.  The plufest is open to members and non-members, and targets interoperability testing of the Ethernet DCB standards-in-progress IEEE P802.1Qbb and IEEE P802.1Qaz (the enhancements required for FCoE). The plugfest will perform multivendor testing across multiple applications such as FCoE, iSCSI, and latency sensitive applications. Participants have until April 20 to register.  Interestingly, the Fibre Channel Industry Association will conducting their second FCoE plugfest at the same venue during the week of May 11. Details here.

White Papers

With 100Gb Ethernet advancing on the standards front, the EA has started to focus on creating white papers addressing 100Gb network deployment. “The Road to 100G: Facing the Challenges of Developing 100 Gbps Platforms” which addresses the challenges of scaling packet networks from the current state-of-the-art, 10 Gbps limited packet-aware, to next generation 100 Gbps all packet-aware, any-to-any networks.  Phew …

This future doesn’t seem very far away.

Cloud Storage at SNW

At this week’s SNW, SNIA announced the creation of the Cloud Storage Technical Working Group (TWG), in a week when cloud was one of the major themes of the show.

The press release got a reasonable amount of coverage on the business wire and various other outlets, so I won’t go into details.  Suffice it to say that the TWG intends to:

• Release a reference model for Cloud Storage with associated terminology definitions

• Solicit proposals for standard interfaces

• Produce a comprehensive set of specifications and to drive the consistency of interface standards across the various cloud storage related efforts.

23 vendors and research organizations have committed to participate in the TWG so far, including NetApp, and the group expects to release its initial cloud service and storage interface definitions in draft form later this year.

The interesting thing here is what wasn’t announced.  For hot new trends and technology, SNIA often announces some kind of marketing initiative -- in SNIA parlance, a Forum (such as the Ethernet Storage Forum) or an Initiative (such as the Solid State Storage Initiative). In the case of cloud storage, SNIA elected to set up a technical group first, one of the goals of which is to come up with a reference model and terminology definitions. 

I believe this is exactly what the industry needs;  cloud is the most nebulous new trend I’ve ever seen in the industry. Once we have a technically robust taxonomy, a marketing organization can add significant value with education and promotion based on a relatively rigorous set of definitions.

The one thing I am concerned about is complexity.  Years ago I worked for an industry luminary (Tom West) who had a theory called the complexity wall.  The theory suggested that for any particular organization there is a limit to the complexity of projects that can be successfully completed.  Projects beyond the organization’s complexity wall will fail.  The theory really does seem to apply in the world of R&D, and I suspect it also applies in standards. 

The SNIA Cloud TWG will need to be very focused to make sure that the complexity of its work doesn’t drift beyond  its complexity wall.  Currently “Cloud” seems to be an ever-expanding universe…

For more information on SNIA's Cloud Storage activities, check out  http://www.snia.org/cloud.

Rechartered SNIA IP Storage Forum becomes Ethernet Storage Forum

At SNW in Orlando this week, the SNIA announced the expansion of the IP Storage Forum to include NAS as well as SAN technologies, and a name change to the Ethernet Storage Forum. The news was picked up by Chris Mellor in The Register and Kevin Komiega in his blog.

Most of the reaction I’ve received while walking around SNW has been, “that seems like a good idea”, but there have been a few questions.  Here’s how I’ve been answering them…

Q:  So, why change?

A: In response to IT, vendor and marketplace pull 

·       We were constantly being asked by IT people why we didn’t cover NAS. Many of them felt that there is a dearth of reliable vendor-neutral information on the topic.

·       We were approached by a group of vendors looking for an industry venue to promote NFS, and particularly pNFS, which is part of the newly ratified NFS v4.1 specification.

·       The ubiquity of Gigabit Ethernet endpoints, the widespread adoption of 10 Gigabit backbones in IT infrastructures, and the rapid growth of Ethernet-connected storage installations was driving an accelerated demand for information covering the whole topic of Ethernet storage.

A broadened focus seemed like a natural succession for the Forum.

Q: Are all Forum members committed to all promoting all the technologies under the “Ethernet Storage” umbrella?

A: Not necessarily.  We have structured the Forum into technology Special Interest Groups: an iSCSI SIG and an NFS SIG initially. Members can sign up for one or more of the SIGs.  The marketing funds available to any particular SIG are proportional to their share of the total Forum members.

Q: Do you have plans to create any other Special Interest Groups?

A: We have received a request to form a SIG focused on the CIFS/SMB protocol, and we are reaching out to that vendor ecosystem for participants.  Hopefully we’ll see enough vendor interest to approve a SIG.  I believe it’s important that we promote and represent the full range of Ethernet storage solutions.  The omission of a large area such as Windows file serving would be a real shame.

Q: Do you have any plans for an FCoE SIG?

A: The Ethernet Storage Forum has no intention of creating SIGs focused on technologies that are specifically promoted by other industry groups. The FCIA is set up to promote standards created within the INCITS T11 technical committee, including Fibre Channel over Ethernet (FCoE).  Our strategy to bring FCoE information into our outreach materials and educational initiatives is via an extension of SNIA’s existing strategic alliance agreement with the FCIA.

Q: Do you have other alliance targets?

A: The Ethernet Alliance promotes the standards created within the IEEE 802.x committees. 

We would like to bring information on higher speed Ethernet (10Gb, 40Gb and 100 Gb) and Ethernet in the Data Center (Data Center Bridging) into our outreach materials and educational initiatives is via a strategic alliance with the Ethernet Alliance.

If you have questions, just let me know.