Announcing SSWG for SFF-8601 Protocol Negotiation for Ethernet Drives

Paul Suhler Paul.Suhler at
Wed Jun 3 15:11:20 PDT 2015

Formatted message: <a href="">HTML-formatted message</a>

This is an announcement for a new SFF Committee specific subject working
group (SSWG).  Instructions for participation are at the bottom of this
SFF-8601 is an SFF Committee standard for transport protocol discovery and
negotiation for Ethernet-attached storage devices.  (It may include devices
using non-Ethernet attachments.)
A number of companies are bringing to market storage devices in a disk drive
form factor which use Ethernet as their primary interface.  These devices are
typically installed in "shelves" which have a backplane connecting the
devices with an Ethernet switch and a controller.  SFF-9639 rev. 0.8
documents the pinouts of two categories of such devices, SNIA Ethernet Drive
and OCP Kinetic.
Discovery and configuration is currently performed in an ad hoc manner.  The
OCP Kinetic specification uses an out-of-band mechanism based on I2C, while
the SNIA Object Drive TWG has been developing a specification for in-band
speed negotiation.
There is a need for short-term and long-term solutions:
1)  Existing designs use single-lane SGMII (serial gigabit media independent
interface) at 1 and 2.5 Gb/s, and possibly faster.  SGMII was devised by
Cisco, who then released their rights; this is not an IEEE standard. 
Moreover, there is no IEEE standard for speed auto-negotiation between those
two speeds.
2)  IEEE 802.3-2012 Clause 73 (Auto-Negotiation for backplane and copper
cable assembly) currently only defines negotiation between 1 and 10 Gb/s,
excluding 2.5 and 5 Gb/s.  Because implementing Clause 73 negotiation would
probably require silicon changes, a change to the standard is necessary for
there to be any hope of switch vendors' implementing it.  A number of
industry consortia are active in this area, including NBaseT, MGBaseT, and
Item 1, the short-term solution, would be the subject of the proposed SFF
standard.  The initial intent is to specify negotiation which can be
implemented without a change in the MAC layer of the transport.
Work Items
1)	Determine scope of work.  This would include:
a)	Transport protocols.  The minimum set would be Ethernet.
b)	Range of speeds (and other characteristics) to negotiate.  The
minimum set for Ethernet would be 1 and 2.5 Gb/s.
c)	 Other subjects as appropriate.
2)	Recruit contributors.
3)	Make recommendations for negotiation protocols for speeds higher than
those covered by the SFF standard.
4)	Prepare the draft standard.
5)	Submit final draft standard to SFF Committee.
a)	The SNIA Object Drive TWG has agreed to let this work proceed in an
b)	Paul Suhler (HGST) has agreed to be the chair of the SSWG.
c)	 Alvin Cox (Seagate) has agreed to be the secretary.
d)	Arnold Jones (SNIA) has offered to set up an e-mail reflector on a
SNIA host.
e)	Upon approval of the SSWG, its work will be announced via various
existing reflectors, and interested parties invited to subscribe to the new
reflector.  A meeting schedule will be established via a Doodle poll.
While only SFF Committee members will be able to vote on the final standard,
participation is solicited from members of the wider storage device and
Ethernet communities.
The first step is to subscribe to the SSWG's e-mail reflector, which is being
hosted by SNIA:
List name: sff8601
To subscribe send an email to
sff8601-subscribe at<mailto:sff8601-subscribe at>.
To unsubscribe send an email to
sff8601-unscribe at<mailto:sff8601-unscribe at>.
To get a list of email commands send an email to
sff8601-help at<mailto:sff8601-help at>.
See the following URL for additional commands:
After sending the subscribe message, you will receive a message to confirm
your subscription.  After you confirm, I will have to approve your
subscription.  You will then receive a welcome message, after which you will
be able to send messages to the reflector.
After a reasonable number of people have subscribed, then I will announce a
Doodle poll to select the time for a weekly teleconference.  I hope that I
will be able to announce the poll by the end of this week.
Thank you very much,
Paul A. Suhler, PhD
Research Staff Member
HGST Research
paul.suhler at
o: 949-757-3432
m: 949-241-6443
3001 Daimler St.
Santa Ana, CA 92705-5812<<a href="">;

More information about the T10 mailing list