FC & IP Made For Each Other? Or Another Shotgun Marriage?

New protocols will test compatibility

The Internet Technology Task Force became the latest battleground in the war to bring IP and Gigabit Ethernet storage networking technology to the data-swamped IT marketplace. Spearheaded by the remarkably active start-up Nishan Systems, the proposed protocols aim to provide interoperability between IP and Fibre Channel based SAN technologies, complement other proposed IP storage technologies, such as iSCSI and FCIP, and might well significantly accelerate the IP Storage industry. Two of the submitted protocols, iSNS and iFCP, are being proposed to become part of a specifications suite within the Internet Protocol Storage (IPS) Working Group.

The fact that Nishan has not yet formally introduced its own product to the marketplace has not put off more than 20 computer, storage software, and data networking vendors who voice their support for these efforts, including Dell, IBM, Nortel, and VERITAS. Other storage industry leaders include ATL, ADIC, BakBone Software, BMC, Chaparral, DataCore, Emulex, Eurologic, JNI, Legato, QLogic, Quantum, Seek Systems, Spectra Logic, StorageNetworks, Sun Microsystems, and XIOtech.

"Submitting these proposed specifications is a significant step forward for the entire storage networking industry," said Aamer Latif, president and chief executive officer of Nishan Systems. "This technology complements the IP storage activities already with in the IPS Working Group and seamlessly integrates with existing Fibre Channel SANs, so we believe these protocols can deliver the IP-based storage networking interoperability that customers demand."

To Bring FC Capability To IP

An overview document summarizing the proposed protocol suite and its architecture has been submitted to IETF. The document states that its mission is to describe an architecture for implementing a fibre channel fabric on an IP network. The architecture is the basis for a protocol suite that provides the set of fibre channel fabric services required by FCP storage devices. The document also includes a philosophical statement of motivation. It states: "The increase in demand for Internet bandwidth has propelled the wide deployment of gigabit networks. This demand has also driven the rapid growth of storage area networks and the development of high-volume IP components capable of meeting the needs of such networks. The purpose of the architecture and protocol suite is to provide the benefits of this IP networking technology to the considerable installed base of storage products and technologies built upon FCP."

The goals of the architecture that Nishan proposes include:

* Enable the implementation of FC fabric capabilities using IP routing and switching elements in place of fibre channel components.

* Provide the fabric services required by storage devices.

* Produce implementations that run at the speed and latency of gigabit IP transports.

* Define the new interfaces to be specified in standards-track or informational documents.

* Identify interfaces defined by existing standards.

The protocols and method of fibre channel frame mapping described in this document permit the transparent attachment of fibre channel storage devices to an IP-based fabric by means of lightweight gateways. This transparency permits implementations to run without modifications.

Bottom line, Nishan and its partners are describing an architecture for the implementation of fibre channel fabric capabilities on public and local IP networks.

Works With All IP Storage Protocols

The protocols address what the proponents feel are pressing needs in the storage networking arena. The first of the two proposed standardstrack protocols to be submitted, iSNS, provides the generic framework for storage entity management in an IP-based storage network. The Internet Storage Name Service (iSNS) is a naming service for IPbased storage networking.

Consolidating existing Fibre Channel and DNS mechanisms, iSNS can leverage standards-based, distributed directory databases, such as the Lightweight Directory Access Protocol (LDAP). The naming service, which functions with all proposed IP storage protocols, including iSCSI and iFCP, has been submitted as a standards-track IETF draft document.

The other key submissions include the Internet Fibre Channel Protocol (iFCP) and Metro Fibre Channel Protocol (mFCP), two complementary protocols for IP-based storage networking. The iFCP links Fibre Channel storage area networks (SANs) and mFCP SANs over any generic TCP/IP network. iFCP is deployable in the wide area or public Internet, as well as in private networks.

It should be emphasized that these are proposals and are likely far from their final form. But the number of players behind the effort can only help promote the effort.

Roger Cummings, SAN Technologist at VERITAS Software, is an experienced veteran in standards-making efforts. Regarding the various architectural efforts, he notes: "VERITAS Software is enthusiastic about Nishan's support for industry standards through its contribution to the IETF. Pursuing this level of functionality is a step in the right direction for the storage networking industry. These proposed protocols are significant elements for enriching storage functionality and supporting interoperable integrated storage area networks using multiple types of interconnect."

The Impact On Tape

If the effort of this group to launch and grow IP SANs is successful, the impact on the tape industry will be profound. Rory Bolt at Quantum/ATL put it well: "Backup and recovery always has been one of the 'killerapps' for storage networking. With iFCP and the other IP Storage elements now being worked on by the IETF, similar backup and recovery solutions in an IP SAN are much closer to marketability."

With all of the various transport protocols in contention, concerns are blossoming how the contenders will impact tape backup. With that in mind, Spectra Logic is planning a tape-over-IP forum that will bring together leaders of the various protocol standards. Scheduled for February, the forum will discuss and demonstrate backup solutions for all of the protocol standards so that backup technology can be built once these standards begin getting adopted.

Such a forum is likely to make a significant sound about preparedness of backup for the numerous new standards proposals, including iSCSI, VI, NDMP, DAFS, and others. It also serves as a wake-up call to vendors and IT management alike. The message: although transport protocols are transparent to end users, they are very much present, very controversial, and need critical professional judgement exercised to leverage the enterprise's IT investment.