Saturday, March 7, 2015

Blueprint -- New Directions For Ethernet: Slower Speeds

by David Fair, Board of Directors, Ethernet Alliance; Unified Networking Marketing Manager, Intel Corporation

Odd title for a networking article, don’t you think?  It’s odd for a couple of reasons, but reasons that reveal the vibrancy of Ethernet.  First, this is a forty-year old technology if you mark its birth from Bob Metcalf’s famous napkin sketch:


What’s remarkable is that this technology even has “new directions” after four decades!  But in fact, arguably, today is the most dynamic period in Ethernet’s history gauged by the number of major new specifications under discussion and development.  The longevity of Ethernet is unprecedented in the tech industry.  The IEEE has completed specs for both 40GbE and 100GbE, the former now in production and the latter shipping today for service provider carriers and sampling for data centers.

And the IEEE is working on a 400GbE specification.  The IEEE also recently added the P802.3by 25Gb/s Ethernet Task Force.  Specific to Ethernet over twisted pairs terminated in the ubiquitous RJ-45 jack, the IEEE has initiated study groups for 25GBASE-T and something called “Next Generation Enterprise Access” BASE-T.  It’s the later I want to talk about.

For four decades, Ethernet advanced on a “powers-of-ten” model from an initial 10 Mbps to 100 to 1GbE to 10GbE.  Part of why that worked was that the ratified IEEE Ethernet speeds kept well ahead of most market requirements.  Bob Metcalf himself started this trend by defining the first Ethernet speed of 2.94 Mbps in 1973, (a rate derived from the Xerox Alto system clock - overkill in 1973 except maybe for Xerox laser printers). To put that in perspective, 1973 was four years before the first Apple II computer became available and eight years before the launch of the first IBM PC.  The “DIX” group (Digital Equipment, Intel, and Xerox) increased the rate to 10Mbps in 1980 in their proposal to the IEEE, who stayed with that rate in the first IEEE Ethernet standard (1983).  But moving an entire Ethernet ecosystem to a new speed is expensive for everyone.  The “powers-of-ten” model helped control those costs.

What changed?  Well, my theory is that Ethernet simply got too successful for the powers-of-ten model.  By that I mean that the volumes got large enough for some specific requirements at more fine-grained speeds to warrant infrastructure upgrades to support those speeds.  And the volumes are large.  My company, Intel, for example, just celebrated shipping our one-billionth Ethernet controller.  We’ve been at it for over three decades, but that’s a lot of Ethernet!  And it continues to grow fast.
Next Generation Enterprise Access BASE-T is a case in point.  The “powers-of-ten” answer to 1GbE 1000BASE-T was naturally 10GBASE-T.  10GBASE-T is now the fastest growing segment of the 10GbE market because it will support up to 100m over low-cost CAT 6A cable and is backward compatible with 1000BASE-T.

But the road to today’s successful 10GBASE-T had a few bumps in it.  Today’s 10GBASE-T PHY is essentially a very sophisticated Digital Signal Processor because that’s what was required to deliver 100m over twisted pair.  And most important for this story is that 10GBASE-T requires CAT 6A cable at a minimum for reliable operation for up to 100m.

The requirement for CAT 6A cable is not a problem for most data centers for two reasons.  First, many existing data centers “future proofed” themselves by installing cable of that quality in their last major upgrade anyway.  And second, twisted pair is inexpensive to purchase and relatively inexpensive to install in an open data center.

It is the rapid growth of wireless access points and increases in their speed specifically that creates the problem leading to a desire Next Generation Enterprise Access BASE-T.   Not in the data center but rather in the office.  Most have built out a wireless infrastructure with CAT 5e or 6 in the ceilings connecting wireless access points at 1GbE, in addition to connecting wired desktops and workstations.  But the latest wireless spec, IEEE 802.11ac can drive bandwidth back on the wire well beyond 1GbE.  And some of those desktops and workstations may be chomping at the bit as well, so to speak, to go faster than 1GbE.

Houston, we have a problem.  And the problem is that ripping and replacing that CAT 5e or 6 cabling in an office environment can be stunningly expensive.  Not for the wire itself but for the installation costs.  So the goal for Next Generation Enterprise Access is to come up with a solution “between” the existing IEEE powers-of-ten solutions.  “Between” in that it has to work on existing CAT 5e or 6 cabling.  And it has to deliver significantly more than 1GbE.  It might try to go to 10GbE if it can, but we know that won’t work on CAT 5e or 6 in general.  If the technology can step down from 10 GbE to one or more intermediate rates it still solves the problem of delivering substantially more than 1GbE on existing cable plant.  And the people interested in Next Generation Enterprise Access firmly believe, based on deployed cable data, that the market size for just this one usage model makes a new spec economically viable.

As often happens in these situations, alliances establish themselves to build momentum to influence the IEEE to consider their proposal.  In this case, there are now two such groups calling themselves the “NBASE-T Alliance” and the “MGBASE-T Alliance” respectively.  Both are proposing intermediate “step-down” speeds of 2.5 Gbps and 5 Gbps, but their proposals differ technically.  The way IEEE 803 speed specifications work is that they define the characteristics of a particular rate and the auto-negotiation processes for recognizing that two communicating devices both support that rate.  So in theory, vendors could develop adapters and switches that support just these rates or even just one of them.  However, the first pre-specification device to reach market supporting these rates also supports 10GBASE-T.  Whether other vendors follow that path is yet to be seen.

Here’s where the Ethernet Alliance brings value to the specification development process by helping to define common market requirements that help the parties come to a common spec in the IEEE.  The Ethernet Alliance believes strongly that the market is best served when “Betamax/VHS” debates are resolved to a common, interoperable solution.  The IEEE 802.3 Ethernet Working Group, after its Call for Interest (CFI) process, established the Next Generation Enterprise Access BASE-T Study Group.  The Ethernet Alliance is advocating that the IEEE 802.3 develop a common specification to successfully address this emerging market requirement.  As Ethernet just continues to grow and become ever more pervasive in new applications, expect to learn of new advanced variants of this venerable technology.  Usually faster, but sometimes slower.

About the Author

David Fair serves on the board of directors of the Ethernet Alliance.  At Intel, David is responsible for driving demand for Intel’s storage over Ethernet (NAS, iSCSI, & FCoE) and RDMA over Ethernet (iWARP) technologies.  He also serves on the board of directors for the Ethernet Storage Forum of SNIA.

About the Ethernet Alliance

The Ethernet Alliance is a global consortium that includes system and component vendors, industry experts, and university and government professionals who are committed to the continued success and expansion of Ethernet technology. The Ethernet Alliance takes Ethernet standards to market by supporting activities that span from incubation of new Ethernet technologies to interoperability demonstrations and education.



Got an idea for a Blueprint column?  We welcome your ideas on next gen network architecture.
See our guidelines.

0 comments:

Post a Comment

See also