IP storage fundamentals
The storage industry is growing quickly, thanks to the much anticipated adoption of iSCSI. This has created a business need to protect investments on costly SANs, and FC devices. The protocols that offered a migration path to IP storage FCIP and iFCP.
Both FCIP, and iFCP carry FC payload and carry the traffic. Both encapsulate FC frames. Both rely on TCP for transport, congestion control/error recovery. But there significant fundamental differences between them.
In the case of FCIP, all the routing is done FSPF(Fabric shortest path first.) All the device discovery, management functionality is done by Fibre channel fabric only (or using SLPv2). Involvement of IP is transparent –as FCIP is a tunneling protocol. Another way to look at FCIP is as the SAN extension technology; Customers who want to extend the distance limitation of their FC- SANs, or those who want to put their storage replicated over long distance for (DR/vaulting) will look for these technologies.
iFCP works much the same way as iSCSI as it relates to routing, security, device discovery and management. iFCP and iSCSI use iSNS(or SLPv2) for device discovery/addressing. (Combination of the SNS service on Fibre channel and DNS of the IP world.) But iFCP end devices are FC-capable only. This is a gateway2gateway protocol as opposed to FCIP which is a tunneling protocol. As IP handles all routing/fabric services, fibre channel fabric is not required. Device2device, device2SAN, SAN2SAN interconnection can all be achieved through iFCP. Then storage routers(Lightsand communications, EMC, McData) can be used to interconnect different protocol implementations as we interconnect Ethernet and Token ring or similar, dis-similar networks....
