Amazon AWS is a Slam Dunk

There’s a battle going on between CEOs and their IT organizations. The CEO is saying “hey – I go home on the weekends, my kids are on Facebook storing pictures and videos for free, Gmail is always on, this new Web stuff is cheap and simple, I can get to these services from any device, Amazon is selling compute and storage for peanuts – why am I spending so much on IT?—Outsource the lot to the cloud!”

IT’s response? “Uh oh – we’re gonna get squeezed. We need to: Virtualize. Simplify. Do more with less. Cut the fat. Increase responsiveness.”

Technology companies, seeing the pickle their best customers are in, the threat to their business and a way to compete, are responding with VMware and Hyper-V integration, thin provisioning, automated tiering, compression, data deduplication; plenty of marketing too – “to the cloud.” And the last year or so has brought lots of high profile M&A, aimed directly at filling portfolio gaps for areas like unstructured data and simplifying IT (Data Domain, 3PAR, Compellent, Isilon, Storwize, Ocarina, etc).  It kind of reminds me of the Three Stooges a little bit – “to the hunt” – lots of action but I can’t help wonder if the big  IT vendors really know where they’re going with this over the long haul.

What I mean is that business is good right now. The market’s up; demand looks solid; everyone seems happy. But there’s a big change coming. We’ll look back five years from now and the gains being made in the data center will be ancient history. CEOs will be happy for a while that CIOs are reducing costs. They’ll keep taking down IT as a percentage of revenue. But CEOs are greedy and we all know they’ll want more; much more. It’s why smart people like Paul Maritz say that VMware needs to move beyond cost cutting into delivering deeper business integration and more substantial value.

There’s a huge gap today between what a service provider like Amazon can deliver with EC2 and S3 and IT’s ability to become “cloud-like.” There’s just too much baggage with that legacy infrastructure and too much complexity. But there are two statistics that stand out:

1. In this decade, storage consumption will grow from 100 exabytes to 35,000  exabytes annually (Source: IDC)
2. Ninety percent of enterprise data that is more than three months old is inactive or never accessed (Source: Wikibon)

The implication – more data generally and specifically more enterprise data should reside and will reside in the cloud at a fraction of the costs relative to traditional on-premise array-based storage. But are IT organizations (ITOs) and their storage vendors responding? I made the case in “The Truth About Enterprise Cloud Adoption” that they’re not responding forcefully and I wonder is it because they don’t get it or because they do get it and it scares the hell out of them?

What Does this Mean for ITOs and their Storage Vendors?

Considering that much of the spending and most of the margin from today’s enterprise storage comes from traditional, physical block or file arrays and the processes built around them, it would seem there’s an enormous disruption coming to the storage world. Why is that? Well, think about how traditional storage arrays are purchased:

• Upfront costs
• Often over-provisioned
• Unused capacity is a depreciating asset
• Software licensing/maintenance model
• Product refreshes every two years or sooner
• A sea of heterogeneous complexity requiring a “human touch” to manage

Sound appealing? Try selling this vision to your CEO.

Nope. The CEO wants pay-by-the-drink, transparent pricing, maintenance included, etc.

Just like Amazon, right? Well, sort of. Amazon has definitely taken once-expensive resources (e.g. compute and storage) and turned them into cheap commodities that come with a monthly bill. Amazon is creating a multi-billion dollar market for billable compute and storage by courting developers, startups and small enterprises. Smart. And big ITOs and their vendors tend to overlook these “niches” – until it’s too late.

For many enterprises, Amazon’s SLAs are not the answer. Amazon’s SLA’s are either too limited or too expensive. EMC’s Joe Tucci characterizes Amazon’s SLAs as “we’ll do our best but if we fail don’t call us.” And CIOs are telling their CEOs that Amazon is too risky, too insecure. Good for the Ruby on Rails crowd but not for our world-class operation.

Well hold on. Earlier this year, Amazon launched a Premium Platinum service for AWS. Why – because it sees a need. And check out the price tag on its enterprise-class service – pricing starts at $15,000 per month! That’s $180,000 a year—enough to hire your own support team. Having sucked in the developer crowd with a simplified rental model, Amazon is turning into a master of the upsell.

Disruption; with a profit.

What are storage companies doing to capture a piece of the cloud storage pie and not get marginalized by Amazon’s rapid growth? The answer is they have to do more than playing the short-term efficiency card and develop cloud-specific, object storage strategies and build a distribution channel with cloud service providers—or in some cases become a cloud service provider.

All the major storage vendors need a platform that offers usage-based billing, secure multi-tenancy, a large scale (i.e. billion plus) object store, support for cloud protocols (e.g. REST and SOAP), a global namespace, proven data integrity and an open set of APIs that allows easy entries and exits into and out of the platform to facilitate integration. AND, very importantly, a channel to sell  these capabilities. These are the core capabilities that will allow the major storage vendors and their ITOs to deliver Amazon-like capabilities and at the same time provide enterprise-class services that can drive premium value for their organizations.

The Vendor Landscape

With these core parameters in mind, let’s look at what some of the larger vendors are doing:

EMC – I wrote in my Predictions 2011 post that Atmos was currently quiet but had mucho potential going forward. That said; I get a lot of backchannel mail from Wikibon users who tell me that Atmos has some issues (i.e. poor performance, challenges scaling to billions of objects, etc). One firm said that it had to do a lot of its own software development to get its Atmos cloud storage infrastructure to work the way EMC promised. Competitors also love to bash Atmos saying it has limited security and questionable support from the EMC field. Here’s what people don’t get. EMC IS IN THE GAME. It understands that there’s a change coming. My bet is over time EMC mashes Atmos up with Isilon and combines the benefits of object storage (cheap, simple, easy access, scale) with a proven scale out NAS management interface. EMC has been clear that it will not compete with its service provider partners like AT&T so look for EMC to aggressively court the CSP channel. EMC is a force that will bring storage business to CSPs through its sales force in an effort to win favor with service providers. This technique works and because EMC is so focused on storage, its sales motions tend not to get lost in corporate politics they way they might at an HP or IBM. Some people have criticized the Atmos business model but I think EMC has it right- sell Atmos as a service provider option and don’t compete with that channel. If it can fix its Atmos product issues, EMC will be in very good shape in this space.

NetApp – Many people forget that NetApp acquired object storage specialist Bycast. In NetApp land, the world revolves around ONTAP so one would think that NetApp will marry Bycast with ONTAP and create an even more unified storage platform. NetApp is getting in the game, but we’ll have to see if its Bycast integration gives NetApp all the capabilities needed for a cloud file system, as outlined above. While it does well in service provider channels it doesn’t have EMC’s muscle and can’t as easily play the leveraged game (i.e. bring CSPs clients in exchange for infrastructure orders).  NetApp is one of those companies that has to succeed based on the merits of its product; how refreshing.

Hitachi Data Systems – Most people don’t realize that HDS has a pretty strong object story. The Hitachi Content Platform (HCP) which leverages the acquired Archivas IP and the “Content Cloud” marketing messages are very good. Last summer, HDS picked up the Parascale IP pieces that were left after the company went under and it may be able to help bring clustered NAS to the existing object platform at HDS. Hitachi has many of the elements, but much like NetApp it needs to integrate them and with cloud service providers bring a cohesive, unified platform story to the market. HDS has many of the cloud capabilities (e.g. usage-based billing and support for cloud protocols) and in many ways has more channel, packaging and market awareness challenges than it does product white spaces.

Oracle – Recently announced the Oracle Cloud File System, a clustered file system freshly stamped with the cloud label. As SiliconAngle’s John Furrier points out, this was epic cloudwashing at its best but as always Oracle is a force and what it doesn’t own, if it’s important, Oracle will buy in an effort to dominate.

HP – Prominently missing from this list so far are HP, IBM and Dell. HP has 3PAR, which is great for cloud compute storage– meaning storage to directly attach to a cloud compute environment. 3PAR itself, however, is not a cloud storage solution, rather, it’s a storage platform that competes directly with the likes of EMC’s VMax and Hitachi’s VSP in the enterprise—both of which also attach to cloud compute installations with a bit more heavy lifting involved. While 3PAR popularized the concept of utility block storage, it’s not an Amazon-like web-based data store. I love 3PAR but it’s not a ‘true cloud’ play. And it seems HP understands  this…

A VAR recently told me that HP is winning business with Nirvanix on top of IBRIX. There’s been no indication of a formal announcement but I fully expect HP and Nirvanix to start going to market together. HP needs to get a stronger unified story and maybe an object cloud play is a better path to leapfrog NetApp’s unified dominance and compete more effectively with EMC Atmos. Earlier this year I said that HP was out to lunch on “Big Data” and it went out and purchased Vertica. Before I heard about this Nirvanix hook up I would have said HP was nowhere on cloud storage but this is clear evidence HP is willing to get serious about object storage and the cloud. Unlike EMC, HP is already competing with service providers (through EDS) and it needs to be more aggressive in this space.  New CEO Apotheker clearly has to play catch up in R&D and acquisitions are proving to be a good use of cash these days. Unlike Rackspace and Iron Mountain, a Nirvanix gives HP a clean play in cloud storage.  Nirvanix offers a competitive cloud file system with enterprise-class capabilities, which when layered on top of IBRIX and HP storage makes for a serious threat to EMC and NetApp.

IBM – IBM is all over the cloud. Cloud will be a huge business for IBM. For IBM however it’s more about analytics, smarter planet, smarter products, etc. From a go-to-market standpoint, IBM ’s Global Technology Services Division is the one to watch in cloud storage. IBM is the gold standard of services companies and that’s where IBM will compete with the likes of EMC’s partners. Notably, when IBM purchased Arsenal Digital a few years ago, it was the services division that absorbed the asset, not the storage group. IBM is still a services-led company and from a cloud perspective that’s a good place to be as it can add value and differentiate from the likes of Oracle and EMC.

Dell is a wildcard. It is in a terrific position to offer cloud storage to SMBs and I could see Dell working with a variety of players. I know Dell partnered up with an enterprise cloud player to win a storage deal at Arizona State University. Dell is another player that has been on an acquisition tear and might just get into the acquisition mix before things get too hot (a la 3PAR). Dell is totally re-inventing itself and storage is a big part of its transformation. Rather than just copying the storage efficiency playbook it may decide to try and get a jump on the competition and disrupt the balance with a true, virtualized, object-based cloud storage offering. Dell could benefit from a fully developed cloud file system layered on Exanet, for example, which would enable it to take on EMC Atmos directly, in a battle for the cloud service providers where Dell can easily bring CSPs business in exchange for an infrastructure buy.

The Bottom Line

The storage cauldron percolated in the past five years, then exploded with innovation and acquisition around storage efficiency—a natural pattern exiting the economic crisis. As we look to the next decade, the shift to on-demand rental models and reduced IT complexity appears to be a permanent change that brings superior business models for most organizations. The consumerization of IT, data growth, analytics in the cloud and the big data trend all point to a new, simplified form of storage infrastructure. Web companies like Google, Amazon and Yahoo have made this model work with innovations like Big Table, MapReduce, S3 and Hadoop, powering Web 2.0 innovations. It’s only a matter of time before this type of simplified technology captures the majority of new data that is exploding in enterprises; and that time is now.