As someone well-placed at a major storage vendor pointed out to me recently, big box storage arrays are going the way of the dinosaur because a single all flash array can replace multiple disk arrays. And for that reason, the all flash revenue potential isn’t anywhere near that of disk even though solid state storage is a hot market. Add to that the fact that storage software can’t make up the shortfall in revenue from exiting disk arrays either and you have a glum near-term outlook for the well-known big box storage vendors. By John Webster
Nevertheless, enterprise storage is still a land of opportunity for those with the right vision. Remember that those who have traditionally predicted meteoric data growth haven’t backed off. IoT is only the latest data volume generator and more will come when we get to some others now appearing on the horizon like blockchain. So there are now some interesting startups looking to capitalize on some hot storage trends. Cases in point include:
Data Protection for a Distributed Computing World
The enterprise is seeing an inrush of clustered/distributed computing platforms for new database and Big Data-driven applications including Cassandra, MongoDB, and Hadoop. For data protection, these generally rely on the creation of three copies of data upon data ingest to affect a recovery when the system is exposed to data loss from drive failure, node failure, data corruption due to operator error and other causes. More traditional backup and recovery processes don’t work in this environment because data consistency across nodes within the system is often a moving target. Unfortunately, users have discovered over time that there are holes in the three copies for data protection strategy. Issues are discussed in this paperby researchers at North Carolina State University. Datos IO—with a new twist on the very mature enterprise backup and recovery opportunity. Datos IO’s RecoverX can now backup and recover distributed, cloud-scale databases at any time interval and at any cluster granularity with data maintained in native formats.
Back in the days of Internet 1.0, a group of UC Berkeley graduate students conceived the idea that the internet could be used as a data storage device of infinite scale. The project designed to realize that vision in 2002 was calledOceanStore. Shortly thereafter, OceanStore was brought to market by a startup named Scale 8 (rotate the 8 in Scale 8 by 90 degrees and you get the symbol for infinity). Unfortunately, Scale 8 went the way of many of the dot com bubblers, but the OceanStore vision lives on. In fact, its thriving. A startup called Hedvig with a software-defined storage that is designed to span any workload, any cloud—public and private—and any storage tier while ClearSky Data offers a fully managed global storage network.
Storage Persistence for Containers
Simply put, containerization in the context of IT is the process of encapsulating an application in a container along with its operating environment. Containers are designed to be flexible enough to run on a range of platforms without dependencies making them highly portable.Docker is currently the containerization platform of choice among enterprise users.
From a storage perspective, the biggest issue with containers is the isolation and persistence of storage. Containerization technology was created originally as a tool for developers that allowed them to streamline application development, especially in cloud environments. Containers were temporary creations as was the stored data associated with them. Now users want to add persistence to the attributes of a container, hence the need for persistent storage. Additional tools such as Flocker (ClusterHQ) provide the storage needed for containers. StorageOS also has a software-defined, full-featured persistent storage platform that supports containers which is currently in beta with general availability expected next quarter.