Customers seeking convergence
The reasons why convergence appeals to storage customers are clear. While there will always be companies who need enormous amounts of storage, most do not need as much as you might expect. The average Fortune 100 company has approximately 22TB of storage distributed throughout their enterprise.
In a previous article, I stated that the key goal for many of these enterprises is the centralization of this storage, and this is still true.
However, a mainstream NAS appliance today can host around 6T Bytes of storage and an enterprise storage array can host around 40T Bytes. Given this comparison, it is clear that in order to consolidate an enterprise's storage, it won't take that many separate devices to achieve the desired level. As a result, it will be the most versatile devices that win in the market, namely those that can supply most of the enterprise's storage needs in a single product or system.
Of course, storage capacity won't be the only key metric in accelerating the convergence of SAN and NAS. Storage services (or virtualization if you prefer) and performance will also be very important factors in the success of this new consolidated market. Good file caching will also be required if files are to be brought back to a central data center from the network edges.
Vendors Rush In
One interesting thing for me in regards to this planned convergence is that the vendors involved in the delivery of these enhanced services form three distinct market segments who are simultaneously moving into this space.
First, you have the storage array vendors, such as EMC, Hitachi Data Systems, and LSI Logic, who have traditionally provided enterprise-class storage into direct-attached or SAN environments. Almost without exception, these companies have added NAS heads to their offerings, allowing NAS services to be provided from their standard storage arrays.
Secondly, there are the NAS vendors who have recognized the need for adding block-level services to their products, initially over Ethernet, using proprietary protocols (although most are now moving to iSCSI) and ultimately offering true Fibre Channel services.
Finally, new entrants to the SAN/NAS convergence party are network infrastructure vendors, including old players like Cisco and newer entrants, such as Pirus Networks, who are attempting to move file-level services onto the network itself, separate and distinct from block-level storage.
This May, I was lucky enough to be offered the chance to chair a session at the JP Morgan Technology Conference, discussing the topic of SAN/NAS convergence. The event was a great opportunity to talk to my peers (and superiors) from all of the largest storage companies (Network Appliance, EMC, LSI Logic, Brocade etc.) and hear their views on this topic. What was made very clear from the session was that no vendor had any doubt that the convergence of SAN and NAS would occur and for them, the main impetus was to increase their market share.
Vendors and market share aside, in a market as competitive as the storage market is today, I think we can soon count on storage features being integrated as a single cohesive solution from a single vendor. NAS and SAN convergence is a certainty, but its form, whether network-centric, storage array-centric or file-centric (NAS), still remains to be decided.
Copyright 2002, Blue Arc Corporation.
This was first published in July 2002