RE: iFCP as an IP Storage Work Item

["Lawrence J. Lamers" <ljlamers@ieee.org>]

The market analysis indicates that a fibre channel over
internet protocol approach, such as FCIP, provides the solution
needed to connect storage networks over Metropolitan and Wide
Area Networks for the foreseeable future. The FCIP proposal
still needs further work in order to provide the necessary
features for it to have long term viability, but we see no
reason that these features can not be added to this approach.

The long term solution  for native attachment of storage
devices to an IP network is iSCSI, a solution that conceptually
fits with the present operating system/software driver stacks,
has broad industry support, and is in process of becoming a
standard within IETF.  ISCSI is a protocol to do SCSI commands
and task management over Ethernet, much as FCP does for fibre
channel, SBP does for 1394, and SIP does for SPI.  Layering one
protocol on top of another to achieve SCSI functionality seems
to complicated an approach that will be difficult to debug and
certify, a field day for Murphy's law.

Certain market niches may be served by other solutions, but it
is believed that the complementary solutions of FCIP and iSCSI
are what should be standardized and promoted to the industry.
This avoids confusion and divisiveness and gets the world to a
storage networked future the fastest,  with the lowest cost and
the least un-productive effort.

It is my belief based on the technical merits and market needs
that the IETF should continue in its present course and support
the FCoverIP work item and the iSCSI work item and not add
additional work items that these solutions already address.

Regards,

Lawrence J. Lamers
Principal Engineer  Advanced Technology Group
SAN Valley Systems, Inc.
408-234-0071
ljlamers@sanvalley.com
ljlamers@ieee.org