Six months after taking leadership of IBM Storage, Ambuj Goyal has declared war on his competitors. “The IT industry is at two major inflection points,” he told SiliconAngle in an exclusive interview at IBM Edge 2013. “I have nothing to lose.”

The inflection points are flash and software-defined storage (SDS) based on open systems, and IBM is going to make a major push in the storage marked based on them.

In his presentation at the opening General Session at Edge he said, “Just a few years back storage was an afterthought for IBM. After we sold everything else, we sold storage. No longer. It’s starting to become a forethought. Data is driving systems architecture now. Processors are just the commodity….”

Different data sets have different values in different businesses, but the economics of all important data is based on the value of continuous operation, not the cost of storing that data. “That’s what the data products and storage products need to deliver, a complete architecture for continuous operations,” he said. “That’s what our enterprise products do.”

The message is clear that under Goyal’s leadership IBM is pushing strongly to regain its place in the storage market. This is good news for users and not so good for the other vendors, who face an invigorated competitor with deep pockets, tremendous resolve. And what it is coming with is flash and open systems.

Flash is the Future

Goyal is betting everything that open systems and flash are the future not just for Tier 1 high-performance storage but for virtually all arrays. While IBM is not close to discontinuing its major disk production operation, Goyal certainly put all the emphasis on flash and never mentioned disk.

IBM already builds large amounts of flash into its System x & z mainframes and PureSystems servers, and unlike some of its traditional large vendor competitors, it uses this flash as primary storage, caching new transactional data directly in flash and then copying it to the array in background. This provides the fastest read/write capabilities available in persistent storage, in the class with flash startups like Fusion-io.

It already has introduced its first all-flash storage array, which it demonstrated at Edge, as well as hybrid flash/disk arrays. And this spring it announced a major initiative to build a series of new service centers worldwide at which its customers can learn the details of getting the most benefit from flash storage in their environments.

In the Edge general session he went beyond that, arguing that, “If you are using 15,000 RPM drives, the tipping point is now. If you’re using 10,000 RPM drives, depending upon the workload, the tipping point is now. And this is not about flash, it is about optimizing systems and software.”

His main argument for this radical statement — one small step from saying that spinning going the way of the old iron donuts as a storage medium —  is that the important issue is not cost per Gbyte of storage but the value of the data, and in particular the value of ensuring continuous business operations in a 24x7x365 global business environment.

“Any time we have a discussion about storage, people talk to me about dollars per gigabytes rather than talking about the cost of data management and how should we do data management,” he said in an interview in theCube from Edge (see below). “If ‘storage’ disappeared from the industry’s vocabulary, then we would focus on data management.”

Open Systems

Goyal also emphasized IBM’s focus on open systems and standards, in particular in Storwize, which he described as third-generation SDS based on the OpenStack initiative.

“We have been doing software-defined storage for almost a decade,” he said. “Our first incarnation was SVC [SAN Volume Controller]. It could leverage IBM and non-IBM hardware, so it was open at a time when there were very few open standards in the industry.

“The second generation software-defined storage is about leveraging open industry standards, so the value created by our partners and our clients holds whether they use our products or not. Our third generation software-defined storage is analytics-driven management.”

Open systems is central to IBM’s strategy not just in storage but across its product lines. “Today we can’t do innovation in everything. There are just too many use cases. The best thing that we can do is to create an environment where our partners and we can innovate together openly and collaboratively. One can create value that the other can leverage that to create new value.”

See the full interview with IBM’s Ambuj Goyal below:

One example of those use cases is a major jet engine manufacturer supplying both the U.S. Air Force and commercial carriers that is using IBM and its partners’ systems to capture comprehensive data on the operations of its engines in flight. The object is to enable more accurate proactive maintenance to head off problems early, improve safety, and increase the percentage of engine up-time. This is part of a new business model in which the engine manufacturer will charge by the amount of time the engine is in operation rather than selling engines using the traditional model.

By building its platform software for maximum openness, IBM hopes to attract as many of the new and established independent tool vendors as possible to build on those platforms in part because their investment then is independent of a single product set or vendor platform. IBM is also enabling its existing hardware on OpenStack, so that any software product written on OpenStack can manage IBM storage hardware.

Data Management Challenges

Several third-party vendors are partnering closely with IBM, providing advanced analytics of various kinds for StorWize. These analytics are vital for the management of data today and in the future. With data coming from multiple sources in very large volume and multiple formats– structured and unstructured– hierarchical, rules-based data management is “running out of steam”.

Today Goyal argues, because Big Data often mandates that the analysis code go to the petabytes of data often involved rather that the reverse, location is often the relevant classification, and in general metadata is becoming much more important.

This is a huge change for many users. IBM Global Services research has shown that “23 percent of our clients do not do capacity planning. Sixty percent do not use virtualization. Seventy percent don’t even have metadata storage.” That presents a huge opportunity for IBM and its partners. It also demonstrates the problem with the concept of storage.

“Storage is such an archaic term. It reminds me of an attic, you throw away stuff in there, you don’t even remember what is in there or what its value is. You don’t even keep records of it, or the metadata. Storage should really be hidden as part of the overall data management discipline. In fact, I couldn’t be happier if the word ‘storage’ disappeared from the industry’s vocabulary. My job would disappear too, but so be it.”

This is becoming an urgent problem with the advent of Big Data. “Everyone knows that data if properly leveraged can create tremendous benefits for our businesses. But the question is, what data do we really need? Data life cycle management techniques which worked in the past are no longer valid in the new world. Should we follow the traditional techniques, FIFO, first in first out? First in, never out, FINO? What are the recovering point objectives for this new class of data? Does it even make sense? The vendors say, ‘If you just buy my storage it will solve all your problems in data.’ Really?

Rather than worry about different storage boxes, the focus should be on overall data management and issues like recovery point objectives and recovery time objectives, that is accomplished in the storage management software layer, and the choice of storage will be part of that overall data management, driven by business needs and informed by comprehensive metadata.

“We think that ultimately the application determines how we want to access data, and that is why we created a third generation of StorWize based on Open Standards.  That’s the model of the future. It is not a box-based or media-based or API-based environment.”