Last year, Hitachi Data Systems (HDS) made an unexpected investment in BlueArc, and also signed a 5-year deal with the company. The attraction? High-performance network-attached storage (NAS).
Prior to the BlueArc deal, HDS had two NAS products, both based on the same technology; a Debian Linux x86 server presenting file-based data through CIFS, via Samba, and NFS protocols.
Andy Williams, HDS' EMEA NAS business manager, said that HDS has a 'file services approach to NAS,' and outlined the product range. One product was a blade inserted in the USP controller. This controlled a set-aside portion of the USP array's block storage for file storage. It was not a NAS head or gateway, instead being, as it were, a NAS file storage reservation inside a block storage silo.
The other product was a modular array, an AMS 200 for example, given a brain transplant, to continue using colourful metaphors. Its normal controller is replaced by the appropriately-packaged NAS x86 board and it becomes a NAS filer, what Williams called: "NAS Attachment for modular storage."
This gave HDS two cuts at the NAS market cake but it was missing a high-end offering, a high-performance NAS product, one up in Isilon's clustered NAS territory. HDS could build one, by developing its existing NAS technology but that would take considerable time and money. Instead it chose to go a combination of the buy-in technology and partner routes by investing an undisclosed amount in BlueArc and signing a deal to OEM its hardware-assisted fast NAS boxes.
These are used for such things as movie rendering, Weta Digital for example, as well as oil and gas data processing, healthcare and financial applications demanding fast access to masses of file-based data.
Williams said that this HPNP, standing for High-Performance NAS Platform, is generating lots of sales activity. In fact: "most of the day-to-day activity is around this." He mentions ATOS IGN as a sample customer, which provides 2D and 3D satellite photo images of France, think Google Earth, to users on the web. The BlueArc kit was brought in to fix slow download problems.
BlueArc has given HDS enterprise NAS credibility. Williams says that, with the BlueArc, HDS can now offer multi-tiered and virtualised file storage and virtualised file servers, complementing the virtualisation facilities in its block storage products.
The HDS:BlueArc deal is for five years, giving HDS time to assess the BlueArc technology, integrate it to an extent in its own technology and strategy, the management console for example, and using HDS arrays as the storage resource behind the BlueArc server, and then push the resulting package to its customers. They can use the BlueArc HPNP as their file services access route with Ethernet access, and the USP-fronted storage behind it as a block storage pool if they wish, with Fibre Channel access.
There is an aspect to this which limits integration of the BlueArc technology with the USP controller line. Since BlueArc uses proprietary and specialised hardware technology it cannot be readily shrunk to a blade form factor and inserted into a USP controller. That may or may not be a disadvantage; it depends how the high performance NAS market develops.
Sun would argue that commodity servers and software will always win out over proprietary storage controller hardware, but then it hasn't got a high-performance NAS product in the BlueArc and Isilon area. The company may be right but it could take a couple of x86 generations before there is commodity HW/SW horsepower available in enough quantity to destroy BlueArc's value.
BlueArc's product strength is made more attractive by NetApp's evident difficulty in entering the high-performance NAS product space with ONTAP GX systems. There is no other OEM for BlueArc apart from HDS and the fact of the investment does hint at the possibility of a full acquisition at some stage.
Whether that takes place will partially depend upon HDS' BlueArc-related sales revenues over the next couple of years, the progress of the high-performance NAS market and HDS' confidence that full control over the technology will give it a decisive advantage. We might as well look into our crystal balls as use any other method to judge if HDS will pay to own the technology underlying its ride into the blue.