Here’s a prediction for 2015: Hitachi Data Systems will make a line of so-called hyper converged infrastructure appliances, the 2U-sized boxes that have server nodes and direct attached storage.
I know this will come to pass because I heard it from HDS’ chief technology officer Hu Yoshida, who was in Toronto on Wednesday meeting customers and giving his technology predictions for the coming year.
One of them was an increase in the adoption of standard converged solutions, which HDS already makes in competition with Hewlett-Packard, Cisco Systems and EMC. “Our fastest growth is in our converged solutions,” he said in an interview.
In fact, he said “I guess most people think of HDS as a storage company, but we are transitioning out of that and we’re more of a converged storage and cloud provider.”
Asked about the smaller hyper-converged systems, Yoshida said “that market is something that needs to be addressed” by his company. “We have a scale-out type file system, but what we need to do is reduce the cost of that and make it more scalable.
To be an infrastructure provider these days a manufacturer has to offer scale up and scale out solutions, he said. “That’s how we’ll differentiate ourselves.”
–A move towards “business-defined IT” because of mega trends including cloud, mobility and big data. Those are the trends that are driving business, and are driving IT. “IT is having to transform itself, not so much about the infrastructure, but more about services. For IT vendors this is also a change for us, because we need to be able to provide IT with the tools so they can be better aligned with the business.”
–The cloud will become a place for Tier IV storage: More file systems and object stores will create interfaces to push things into the cloud, “because cloud is the cheapest storage you can get.” Public clouds will be used more as repository for data enterprises rarely retrieve. IT will still control the management of the metadata and encryption from the data centre. Where the cloud becomes expensive is when you retrieve data, because you get charged.
— Greater focus on data recovery tools. Organizations can’t afford to have 60 copies of a database, so there needs to be better data management software, including better visibility tools so users can see there are multiple copies of things.
–More use of “global virtualization,” meaning virtualization across multiple servers, so IT can disperse storage across distances but have one global image. Advantage: data can be move access closer to end user, reduce migration across distances. Already available from a number of vendors.
–More intelligence in enterprise flash drive modules to increase capacity
–You’ve heard of big data. Get ready for data lakes, huge repositories (exabytes) of raw data where you bring apps and process what you need.
— More management automation through templates:”Push a button and everuthing gets provisioned.”
–More “software defined” everything. He admits that there is hype around “SDx,” but believes there’s value in the idea of software defining the value of (fill in the blank).
–More analytics to make sense of big data and the Internet of Things.