On March 3rd, HGST announced an agreement to acquire Amplidata, a privately-held developer of object storage software. Here is the official announcement of HGST acquiring Amplidata.
I, for one, am a fan of object storage and Amplidata’s technology as well. Himalaya, Amplidata’s object storage offering, is a software–defined storage solution designed for multi-petabyte environments. The technology leverages erasure coding to provide high levels of resiliency for massive capacity environments, while leveraging commodity hardware for greater affordability. While I currently do not have any information on the cost of the acquisition, based on what I do know, I applaud the move for Western Digital and HGST.
The move is interesting on several fronts, and I wanted to take some time to discuss a few of these.
- Ever increasing pool of storage players: In my storage predictions for 2015, I mentioned that I expected a number of technology providers, who may have been considered to be component suppliers, to be beefing up their storage solutions offerings. New innovations in storage technology are changing the storage market landscape and opening up more opportunities to compete. Also, the potential of vertical integration to help make storage more affordable seems to be too enticing for many technology providers to stay on the sidelines. By adding an object storage offering, HGST is able to provide a richer suite of storage solutions to enterprise customers.
- End-to-end cloud storage: HGST may have put together one of the more interesting cloud storage technology portfolios. In December of 2014, HGST completed an acquisition of Skyera, a provider of solid-state storage. Skyera’s technology offers high-performance solid-state storage in highly efficient dense building blocks serving transactional workloads. Amplidata serves the other end of the storage spectrum, providing affordable and massive scale infrastructure for unstructured content storage with support for the S3 cloud storage protocol. The combined portfolio enables HGST to offer architectures optimized for both ends of the workload spectrum, high performance or high capacity, along with HGST’s existing technology from previous acquisitions to provide server-side solid-state caching. Time will tell if HGST decides to keep these solutions separate or bring them together as a more integrated cloud storage solution.
- Storage defined by workload rather than protocol: I have been suggesting for a while now that storage definitions will eventually have to change. The current market segmentation by protocols, such as Fibre Channel, iSCSI, or NAS, make less sense given the emergence of solid-state object storage technologies. For high performance transactional workloads, architectures that optimize solid-state are going to have far more impact on infrastructure performance and economics than the network protocol. On the other side of the coin, object storage provides augmented protection for massive capacity environments along with cost optimization for on-premises cloud storage environments that can rival or surpass public cloud storage. The bottom line is that if storage moves in the direction where organizations shift away from general purpose storage arrays to an on-premises cloud environment with back-end defined by a blend of performance-optimized and capacity-optimized storage architectures, HGST now has both sides covered. And if HGST is able to combine these technologies into a more integrated solution, then that would really be interesting.
As always, I am interested in your thoughts and comments. Disagree with my thinking? Please let me know.