IBM mainframe cloud innovation. Worst oxymoron evar? Or …

December 5, 2013
    Even Ken Jennings, the longest winning streak on Jeopardy!, knows Watson is a real innovation

Even Ken Jennings, the longest winning streak on Jeopardy!, knows Watson is a real innovation

There was much shared mirth on Twitter recently over a blog post by Quentin Hardy (@qhardy), the Deputy Tech Editor and The New York Times.

The short post was titled IBM’s Big Plans for Cloud Computing, and it is worth a read, if nothing else than for a little context of what IBM is thinking about cloud.

Featuring Lance Crosby (@lavosby), CEO and Founder of SoftLayer (recently acquired by IBM), it did indeed have some choice hyperbole.

For example, this gem:

In 2014, the company will make a series of announcements that will shiver all challengers … “It will take Amazon 10 years to build all of this,” [Crosby] said. “People will be creating businesses with this that we can only dream about.”.

Whenever IBM in general, and mainframe in particular comes up on Twitter, there is a disbelief in the connection of the words ‘mainframe, ‘cloud’, and ‘innovation’. Yet already enterprise customers are usign mainframe clouds, and IBM is talking about mainframe as a public cloud service. I think by any definition, and certainly by my own, these are  innovations in cloud. But many disagree.

Dell’s James Urquhart (@jamesurquhart) and I went back and forth a little about this when inevitably the conversation turned to the persistent myth that IBM cannot be believed prima facie as an innovator in cloud:

In a surprise to no one, George Reese (@GeorgeReese), the Chief Always-the-right-room-for-an-argument Officer of Dell’s Enstratius team, piled on:

Not to pick on George or James – they are both very smart people with great insights and ideas – but rather to use their posts as exemplars of a common (mis)perception, this is again the myth that large companies cannot innovate. This variant – that ‘large legacy vendors cannot innovate’ (deliciously ironic coming from Dell) – is no less incorrect. I’ve a whole thesis on this, but let’s stick with IBM for now.

Sure, we can fairly say that IBM’s discontinued (is ‘failed’ too strong?) Smart Cloud Enterprise trailed Amazon’s innovation in cloud computing. And even despite the article’s claim of $1bn in cloud revenues, major investment in OpenStack, and now a major cloud service provider in Softlayer, maybe you can still say IBM trails Amazon in cloud innovation.

But you must realize that this is comparing just one or two of IBM’s many cloud offerings, and just one relatively small part of IBM’s overall business, with AWS’ entire business and its complete reason for being. Then you just need to do the math. How many new features is AWS producing each year for all of its research and products? How many new features is IBM producing for all of its research and products? Even for just the 140+ solutions it plans to make available via the cloud, or the products responsible for its billion dollar cloud revenue?

However, maybe this is not innovative enough for some. Fair enough. It is only ‘incremental innovation’, which I know a lot of people discount from ‘disruptive innovation’.

Consider then, a Watson-enabled mainframe cloud, perhaps hosted by Softlayer as a pure IaaS or PaaS offering.

In which case, please consider Watson, the POWER-7 based supercomputer, one of the most innovative and powerful large-scale computing developments of this millennium.

Even if Watson is not a mainframe per se, consider it in context of the multi-architecture nature of modern zEnterprise mainframes, which already run z/OS, Linux, POWER, and x86 (Windows, Linux) in a single architecture. Combining these systems could host and execute Java, C++, Cobol, PL/1, Ruby, Python, MySQL, Assembler, Prolog, Hadoop, CICS, IMS, DB2 and more, all in a single directly-connected architecture.

Consider then, a Watson-enabled mainframe cloud, perhaps hosted by Softlayer as a pure IaaS or PaaS offering. Yes, ‘true cloud’ again. This could be a massively parallel, high-performance system with in-memory storage, enterprise throughput, unmatched reliability, and a “natural language processing, machine learning, and hypothesis generation and evaluation to give you direct, confidence-based responses” – all on-demand, infinitely scalable, and pay-per-use.

Then consider, as Crosby notes, that IBM also are planning to bring to their cloud offering a unique experience in a wholly different realm of cloud, specifically:

the insights it has gained from building and licensing technology used by Microsoft in the Xbox gaming console, and Google in its own network operations

This can then start to build on knowledge of MMO communities, network graphics delivery, principles of gamification, visual and multimedia analysis, deep web crawling, semantic search, and much more.

Now, imagine even just the analytics applications of this capability, on this architecture, like:

  • Correlating social campaigns against credit card, PayPal, Bitcoin, and EFTPOS transactions
  • Analyzing real-time voice and data from mobile devices for sentiment analysis  … and response
  • Producing demand analysis – and prediction – from live forum chatter ahead of product launch

The big social data applications alone are potentially world-changing. Not to mention cloud-migrant versions of existing Watson use cases in areas like healthcare and finance. And this is just me spit-balling – imagine the ideas that the best of the best IBMers are coming up with!

This is barely even in the future. IBM is already opening up Watson “as a development platform in the Cloud“. Partners are building apps and IBM is building out its marketplace. IBM zEnterprise BladeCenter Extension (zBX) already runs Watson-compatible POWER7 processors. It not a far stretch of the imagination to think IBM might integrate these capabilities and provide a new cloud service – one that would enable game-changing business innovation that would make Amazon look like ‘your mess for less’.

So, is IBM 10 years ahead of Amazon in cloud? In innovation? There are good arguments both ways, and 10 years is a long time in technology. But do not dismiss out-of-hand IBM’s ability to innovate with cloud, or anywhere else. I do not know what exactly IBM is planning, but there is no lack of evidence for IBM’s actual innovation, and their potential for innovation, in the cloud and beyond.

Tags: , , , , , , , ,

3 Responses to IBM mainframe cloud innovation. Worst oxymoron evar? Or …

  1. Dave
    December 12, 2013 at 07:57

    My thought is, “Who cares what’s behind the curtain?” Most cloud customers want the hallmarks of cloud and could care less what iron it’s running on. Can I get it provisioned quickly? Can I scale up and down? What flexibility do I have in OS, configurations? What features are available to me? What does it cost?

    Those benefits will outweigh what is powering the vendor’s cloud, be it 100,000 white box x86 machines or 1000 mainframes.

    • December 12, 2013 at 21:48

      Dave, thanks for reading and for your comment.

      In part I do agree – the focus should be on the utility (literally) of the offering, the business value of the service. The platform it runs on is definitely secondary.

      That said, the mainframe offers certain utility and business value specifically because of the platform. The availability, throughput, data access, scalability, and more are key differentiators of the platform that create unique business value that any number of x86 machines do not.

      And that is without considering certain applications that only run on a mainframe – and their unique business value.

      In an ideal world, the iron would not matter. But until all workloads are perfectly portable between environments, it still does – for better or worse.

      Thanks again for an interesting viewpoint.

      Andi.