Earlier this week I explained what software-defined storage (SDS) or software-led storage is. In this software-defined storage world there are some important big players that are currently leading the space, and on these four backs we’ll ride into the first iteration of software-defined storage. Hewlett-Packard, EMC (who’s growing federation of entities include VMWare and Pivotal), NetApp and Fusion-io are the ‘Fab Four’ in SDS. All four are making huge strides to move the meter in SDS, where storage is included in the software that is provided across the data center.

In this piece I’ll review these four major market players, outlining their current offerings, strategies and market positions within the SDS space.

Recently, Wikibon published a research note defining what Software-Led Infrastructure, and storage is a key component of that architecture. Here is an excerpt from David Floyer, CTO of Wikibon explaining software-led storage:

“The fundamental premise is that data centers will migrate from the current model where all storage services are provided from software within the specific “box” component (e.g., a storage array) to an infrastructure where the system, storage, and network services are provided as software across the data center as a whole, and the software within each component is much reduced.”

HP Moves Towards Software-Led Storage

Hewlett-Packard announced a large number of products and product updates at HP Discover 2012 in Frankfurt, Germany and began tying into the software-defined theme. While it was clear that a fundamental shift was beginning at HP, moving from a hardware to a software-led player, the “platform story” (i.e. the ability to connect to a horizontal platform through open APIs) didn’t come through.  From that event, Wikbon identified StoreAll as the most strategic and innovative announcement out of HP Discover 2012. StoreALL is a potential game-changer in the management of unstructured and semi-structured Big Data. An assumption of software-led storage is that by 2015 a strong majority of new arrays for active data will be flash-only. Supporting that premise, HP introduced SSD-only 7000-range arrays.

Another sign is that the HP 3PAR Array Operating System will include a Web Services Toolkit and Representational State Transfer (REST) API to allow the integration with software-led infrastructure automation tools using a standard interface. HP’s 3PAR system is a converged infrastructure offering, and HP’s lead horse in the software-led storage race.

Keys for HP: move from forcing as much software as possible to be consumed at the box level to a “play nice” or “open and choose” strategy, where services and boxes can interact through stable APIs. But the real action from HP will come from OpenStack initiatives. Cinder (block) and Swift (file) are open source instantiations of software-led infrastructure and HP is betting its cloud strategy on OpenStack. One thing observers should watch is the extent to which HP’s storage group, under the leadership of GM David Scott, will contribute to OpenStack and “plug in” to the framework.

The EMC, VMWare, Pivotal Federation Play 

EMC showed its teeth in February 2012 when it announced VFCache, ‘Project Lightning’. It was the first server-based flash storage card system from a major server or storage vendor; but it was behind the leader Fusion-io. Nonetheless, as we explained earlier this week, flash is going to be a major player in software-led storage, so a large vendor acknowledging that fact in this way should not be taken lightly. VFCache can work with any vendor’s storage, and included in the announcement EMC showcased Fully Automated Storage Tiering (FAST) software, which optimizes the placement of data on the most appropriate device (from flash to spinning disk). FAST is a fundamental part of EMC’s flash strategy in the near-to-mid term although many question the need for tiering longer term.

Because it’s organic offering wasn’t going to cut it in the marketplace, EMC had to make a move and bought XtremeIO, an all-flash array player out of Israel. The company is reported to have excellent technology but it’s taken EMC a year to make it ready for its demanding customer base. Earlier this year, EMC made a big splash with XtremeIO, announcing both all-flash arrays and server-side software to work on specific cards. Reportedly, EMC OEM’s the Virident card through Seagate and delivers software on top to its customers. The product is not quite ready for prime time and is only available to selected customers.

Nonetheless, EMC is expected to dominate in this space. EMC is the market leader in storage and majority owner in VMWare, the market leader in server virtualization. So EMC not only has significant market-share and a head start, but it has significant skin in the game, and thus skin to lose. EMC is looking to make a very strong attack on the software-defined data center (SDDC), and to disrupt the software-defined space as a whole, they are trying to put forth a vision which involves a federation of companies. The Federation strategy is 3 parts:

1. EMC’s focus is on information infrastructure
   1. XtremIO all-flash array and update to server flash boards (now named XtremSF)
   2. ViPR, EMC’s software-defined storage platform
2. VMWare trimmed down its strategy and is doubling-down on solutions that virtualize, pool, and automate the data center
   1. App-store for the enterprise is gone, Cloud Foundry, Gemfire, Spring and Cetas are now part of Pivotal
   2. Laser-focused execution on Software Defined Data Center, Hybrid Cloud, PC to Mobility (post-PC era)
3. Pivotal is a platform play, smack at the middle of the Big Data and Cloud intersection
   1. The new platform: Big Data, Fast Data and Rapid Application Development sitting on a cloud abstraction and application automation layer
   2. Technology shifts and growth of service providers and cloud-based services define a new relationship with the platform

Keys for EMC: EMC’s ViPR was announced as a “platform.” It’s how EMC intends to attack the OpenStack and open source movement. EMC’s philosophy is you don’t have to be open source to be open and it will bundle proprietary function into its platform. The key for EMC will be to demonstrate that its closed IP can be accessed through open interfaces and, very importantly, add value beyond what the open source crowd can deliver. EMC has to walk a fine line. On the one hand the platform play can expand its total available market (TAM). On the other hand, EMC risks cannibalizing its legacy products (e.g. Symmetrix and VNX). EMC has publicly said it would rather eat it’s own lunch before someone else steals it. Also, don’t forget that VMware made an an acquisition of Virsto recently which puts it into the software-defined storage mix. The interrelationship between EMC and VMware will likely play a big part in EMC’s strategy.

NetApp Looks to Offer Best Scale-Out Software-Defined Storage Solution With ONTAP

NetApp has succeeded thus far by identifying the right trends (e.g. virtualization), and capitalizing on them, while choosing the right partners along the way. The next big wave NetApp is tackling is clustering. They basically need to manage the massive installed base of filers by letting its products look more like a single system via a clustered, scale out model. This will simplify hardware + software for seamless and efficient scaling of storage. Scale-out is challenging technically and while they’ve launched some pretty disruptive products in the past, their storage solution has had trouble scaling.

Enter ONTAP.  NetApp is betting the company on Clustered ONTAP, a clustered offering that allows their storage solutions to scale-out; and be managed as a single entity. The architecture of ONTAP places the active, frequented or accessed data on flash and puts the rest on SATA drives. It solves a sophisticated problem of complex computing systems, and ONTAP is years in the making. The downside to this extensive time spent on R&D is that it left competitors with more time to attack. NetApp plain and simply thinks it can outpace EMC, and its CEO Tom Georgens isn’t shy in telling anyone who will listen; now it’s time to demonstrate that.

Keys for NetApp: NetApp is much smaller than its larger competitors (EMC, IBM, HP, Oracle) and is not in as strong a position to build a large and powerful ecosystem. At the same time, NetApp should be able to innovate faster than its larger competitors. Herein lies the challenge for NetApp as it relates to Software-Defined Storage. If the world is moving toward platforms, developers and partners will prioritize those that can drive the most amount of business. NetApp must demonstrate that it has enough market “juice” to attract the ecosystem and get them writing to its SDS platform.

Fusion-io Aims to Win With Hyperscale Markets

It is the belief of analysts at Wikibon that Fusion-io is going after the hyperscale market. Fusion-io’s ability to do atomic writes, combined with its experience in hyperscale, make it the clear leader in that market. With its main competition coming from EMC, Wikibon CTO David Floyer believes that Fusion is better positioned for hyperscale markets, and EMC’s competitive benchmarks were misleading. He’s said that while EMC may have faster cards for traditional PCIe connect, Fusion has roughly a 2-3 year advantage over EMC and others with atomic writes (i.e. the ability to write directly to flash and bypass the disk protocol). Many in the industry are banking on NVM Express, an industry standard to do atomic writes that has been proposed but is moving slowly through committees. There is no secret that the high-end database performance is where Fusion-io shines, with its rich set of software to optimize database performance.

A recent bit of news that rattled Fusion-io was the departure of former CEO David Flynn and CMO Rick White both cofounders; to which the newly appointed CEO Shane Robinson said in an exclusive interview with SiliconANGLE that the company was not for sale and still very stable.

Keys for Fusion-io: It’s a forgone conclusion that Fusion-io’s success depends largely on how fast it can enlist ISVs to write to its SDK, thereby locking in an ecosystem advantage. Wikibon Co-Founder and Chief Analyst Dave Vellante believes that the big data, NoSQL and open source database crowd – e.g. MySQL, Aerospike, HBase, Cassandra, Mongo, Accumulo, etc. will hop on Fusion’s approach and pressure Microsoft and even Oracle to exploit Fusion-io’s method. Another key for Fusion will be the transition to its new executive management. It’s unclear if this will be a distraction or a boost for the company. Thus far it’s been a distraction as the stock reacted negatively to the news on the assumption there was something wrong with the company. Wikibon’s Vellante believes this is overblown and told me “Fusion-io’s demand comes from the hyperscale guys and often it’s lumpy. The big hyperscale guys buy in huge quantities and then stop buying and then start again. I’ve seen this cycle with Fusion before and would be surprised not to see it again before the end of this year. Fusion still has a big lead in the hyperscale market.”

Whoopty Do, What Does It All Mean Basil?

Calling Big Data…well Big Data doesn’t come close to doing it justice. That’d be like calling Goliath in David vs. Goliath, average height. Software-led storage is currently being targeted as a $50 billion industry. But I think the collection of software-defined verticals plus Big Data is some ten or eleven iterations bigger than that. Technologies haven’t even been invented yet that three years from now are going to be ‘the norm’.

The software-defined storage (SDS) market may be led by the so-called ‘Fab Four’ but I liken it to a checkers game. Most people don’t have a specific plan going into checkers. The overwhelming consensus of people play checkers defensively until they catch a big break. Back-to-back-to-back jumps, a solid double-jump, and then quickly they are the first to utter the magic words, “king me.” What’s more there are others waiting in the wings. IBM in particular hasn’t weighed in with its software-defined strategy; neither has Oracle; and there are a number of smaller players (e.g. Gridstore, Datacore, Sanbolic, etc.) that are truly defining software-defined…the problem is no one’s ever heard of them and they have no megaphone. But they are clearly acquisition candidates that could supercharge someone’s business.

The board is just seeing its first moves. The real action will start when we see the first jump. We’re a long way away from anyone, ‘Fab Four’ included, saying “king me.”