BANGALORE, INDIA: Chandranshu Singh, senior analyst at Ovum India explains what will be the storage protocol that the market will embrace as it moves forward.

CIOL: As more and more data centers migrate to 10GbE, which protocol – FCoE or iSCSI - has the most to gain or lose? Why?

Chandranshu Singh: FCoE is an attempt to keep the FC market alive; it makes sense for those organisations that have significant existing investments in FC technology.

Also Read: Can iSCSI win over FCoEs hype?

iSCSI is a superior technology, and many benchmarking tests have lain to rest the discussion about iSCSi’s performance and network latency. Furthermore, iSCSI is cheaper to implement at the hardware level, so it makes more sense for organisations that are now looking to invest in storage networking technology and have no fibre channel legacy.

CIOL: Do you think iSCSI/FCoE is an 'or' (iSCSI or FCoE) or 'and' (iSCSI and FCoE) situation?

Chandranshu: At present, it is an ‘or’ situation. Businesses with legacy FC investments can protect and extend their investment by running FC over Ethernet.

At the same time, organizations looking to deploy storage networking are looking towards iSCSI as a cheaper and better alternative. In due course, when organizations go in for the next network refresh, we expect to see iSCSI (or perhaps some other disruptive technology) replace FC and FCoE SAN environments.

CIOL: Do you see a trend of multi-protocol storage networking in data centers gaining prominence, especially in India? Why and how?

Chandranshu: From the supplier side, multi protocol (or ‘unified’) storage devices have been available for a few years now. On the consumer side, organisations have to identify their storage needs and choose a technology that meets their current requirements and can be scaled up to incorporate additional traffic in the future.

Also Read: iSCSI sorts out performance issues

Small organisations that need a ‘generic’ storage infrastructure that can adapt to block as well as file-based storage requirements will benefit most from a solution that combines iSCSI or FC/FCoE with NAS.

And it doesn’t require a multi-million dollar investment. So the benefits are there for people to see: simplified management, less resource intensive, incremental deployment, and cheaper. In our view, this trend will catch on with small and mid-market businesses in India.

For larger enterprises, such procurement decisions are typically made at the department (or line of business) level because storage needs are application/process specific.

However, today’s technology is mature enough to support high I/O volume (typical of enterprise data centres) and the apparent benefits aren’t lost on them. We expect the adoption to be slow and sporadic in the larger enterprise segment.

CIOL: Which were the factors that prevented iSCSI from extending its reach into fibre channel’s domain?

Chandranshu: The only credible argument against iSCSI was Ethernet bandwidth, now with 10GbE Ethernet, an iSCSI SAN looks more promising (and cheaper) than FC or FCoE.

Also Read: Time to get over the fibre fixation

Another related argument against iSCSI is that of TCP overhead; in our opinion, the overhead added by the larger sized TCP/IP headers (compared to FCoE packets over Ethernet) is nominal compared to the disk latency and thus inconsequential.

As things stand, there is nothing that prevents iSCSI from breaking into the FC domain, it will likely happen with time.