ProphetStor was founded by experienced storage experts circa 2012 and over the years, the company has been committed to introducing innovative solutions for bringing intelligence into Software-Defined Data Centres (SDDC) and in 2018, it seems like they’re about to step up their game even further and prove that they’re more than just another storage or BC/DR vendor. DSA was given the opportunity to interview the company’s founder and CEO, Eric Chen, when he was in KL and the first thing we noticed upon seeing his business card was that it said ProphetStor Data Services instead of, say, ProphetStor Storage.
Speaking about the inception of the company and why the company was named that way, Eric said, “We don’t use ProphetStor in a religious sense at all. In the past, when we talk about software defined storage (SDS), we didn’t just mean putting a piece of software on a commodity server and turning that into traditional storage. We believe that we should be doing more. We believe that storage itself shouldn’t be a passive device. It should play a central role in the whole computing environment and what we want to bring to the industry is the next generation of storage.”
The need for intelligent data services
And that’s where the “data services” part comes in. ProphetStor’s mission, according to Eric, is to deliver intelligent data services for applications and when storage becomes intelligent, proactive, flexible, adaptive and predictive, it would have much more value than just capacity. He believes that the storage services should understand the application in order to support it according to its needs, be it IOPS, capacity or other requirements and provide it just in time – optimising all available resources. But in order to do that, it’s important to be able to predict how the applications would behave in the future.
“When we understand what the application needs, there’s a lot that we can do in storage. If we can have a way of predicting the behaviour of applications as well as the infrastructures, maybe understanding the failure patterns, understanding the reliability of certain items, how you combine them even to provide better reliability, there’s all kinds of work that we can do. And in that case, we actually have a very flexible, adaptive and also intelligent environment for serving the applications. Thus, we would like to bring AI or machine learning technologies into the driver’s seat of managing all these types of resources,” Eric explained.
As technologies like virtualisation, cloud reach maturity and adoption of emerging ones, such as containers, continue to increase, the complexity within the data centre also increases tremendously, making it almost impossible for human system administrators to efficiently allocate resources, let alone predict what’s going to happen in the future.
For these reasons, the addition of intelligence through AI has become a necessity rather than a luxury to get storage services to where it needs to be in the SDDC era. To lay the foundation to enable ProphetStor to move forward, Eric's team of engineers worked on securing various patents to strengthen the company's global patent portfolio. Just this year, they have been assigned several new patents to make storage more adaptive, leverage AI-enabled predictives to optimise resources and solidify ProphetStor’s storage agnostic data services vision.
Real world case study
As a testament to ProphetStor’s commitment to innovation and challenging the status quo of storage, Eric explained about a case study that they had done with Sumitomo CSK (SCSK), one of the largest cloud service providers in Japan. SCSK wanted to move away from conventional storage, which they found to be expensive and inflexible, to SDS. Eric pointed out how difficult it is to gain the trust of tier-1 service providers in Japan, but by showing how ProphetStor’s solutions can be adaptive, with intelligence built in, SCSK started engaging with ProphetStor 18 months ago. Further details on how ProphetStor was able to meet SCSK’s demanding system management and budget requirements can be found here.
ProphetStor has also been working closely with Intel to validate the claims that its patented AI-technology built into DiskProphet is able to predict disk failures with up to more than 95 percent accuracy. Eric mentioned that this ability to predict disk failures is absolutely crucial because disk, be it spindle or SSD, makes up the foundation of an array or hyperconverged infrastructure.
He continued, “By predicting the life expectancy, you remove the uncertainty of handling the random errors and failures of the disks, and turn that into a scheduled maintenance. Intel actually provided those servers to us so that we can implement our AI engines and also the models in there for a long period of time to validate our solutions. So Intel has given us their mark of approval, and with DiskProphet, we have been working with many storage companies, warehouses, high performance computing and vSan environments. We’re also working with the Ceph community.”
ProphetStor’s overall objective for all this, Eric said, is to bridge the gap (in terms of expected reliability, performance and availability, etc) that currently exists between the incumbent storage and software-defined offerings available today. And with artificial intelligence, customers would be able to explore and exploit ideal resources and turn unused resources into good use.
On the local front, ProphetStor will be looking to move into the area of AI for the whole ASEAN region in 2018 and beyond due to the growing demands for the technology. The company will work on addressing this need for AI first through hardware failure prediction capabilities before moving into workload analytics and eventually, the AI operations of the entire data centre.