Friday, August 6, 2010

Sea Changes, aplenty!

I caught up with a good friend in NonStop, as I was in Cupertino recently, and a casual remark that he made to me has aroused my curiosity. In isolation, it may have little meaning, but in terms of the bigger picture for IT, I am beginning to wonder whether we are witnessing a sea-change or even the end of an era. “We continue to ship NonStop systems but the number of really big systems we ship has diminished in recent times,” my friend had remarked.

I know there are still a number of large NonStop deployments out there – in the telco and messaging marketplaces, and more recently, in secretive homeland security roles as well. But with the external circumstances including industry mandates and government legislation, the emergence of secondary, tertiary, and beyond data centers, what is being installed in these multi-site deployments are systems typically a lot smaller than in the past.

In my most recent post to Real Time View, "Measuring up?", I closed with comments lifted from the paper “the coming of Porous Enterprise” edited by Mark Plakias, VP Strategy, Orange Labs SF, which is a Californian branch of France Telecom R&D. In that post to Real Time View, I included a reference to the observation “what seems to be happening is a shift not only from information to knowledge but from knowledge to creativity. Indeed, given the power of networks, knowledge is slowly becoming commodity.”

However, also contained in the paper is another very interesting observation of how business, or as the paper calls it, “the Firm”, “as we have known it … is shape-shifting, re-centering around a new functional model of what is ‘core’ and opening its boundaries into a more porous way to better connect and engage with competitors who want to collaborate, communities that require accountability, consumers who expect innovation, employees who demand authenticity and meaning, and technology that waits for no one.”

Finally, one other quote also caught my attention as the writer asks “why do we sequester ourselves in big buildings when more and more of the information we need is outside in the world of customers, partners, and a growing collective pool of knowledge on the web?”

From the very first time I heard HP’s message of Converge! Transform! Innovate! I wondered about the final step – the jump to innovation. From what I have read of late, I have become more aware that today, for those businesses that successfully navigate from convergence through to transformation, there’s the likely possibility that creativity will flourish as a result. It’s only when creativity begins to emerge that innovation will truly surface.

We cannot develop the creativity behind the walls of large corporate edifices – look at all the papers (check out the July – August 2010 issue of Harvard Business Review and the article on “Stop the Innovation Wars”) bemoaning headquarters’ establishment of innovation departments and the disappointing results that so often follow. Rather, dispersing corporations to where employees work more closely with their partners, customers, and potentially, even competitors, in a decentralized manner (rather then a regional or branch distributed model of the past), is the theme of this paper.

“Decentralization + Creativity = Porous” is the conclusion that appears late in the paper and it really halted me in my tracks. When viewed alongside the earlier comment “more and more of the information we need is outside in the world of customers, partners, and a growing collective pool of knowledge on the web?” there’s little argument to be made against today’s trend of broadly connecting everyone in a business with everyone else in the businesses ecosystem.

The slowing in demand for big systems, apart from select specialized data base deployments perhaps, may be in fact a reflection of the “shape-shifting, re-centering around a new functional model of what is ‘core’” that we are all witnessing of late? Have you ever wondered why so many public companies have decimated the ranks of their junior executives / senior managers? And for most of them, the business hasn’t missed a beat as stripping away the layers has brought more people into direct contact with all elements of the business world? Smaller, empowered, groups focused on newly-defined “core” components of the business are demanding more and more system nodes in support of the transactions and data that are crucial for them to perform the roles now assigned to them.

Again, not a return to distributed models of the past, that were by definition hierarchical as information flowed slowly up the ranks, but more of a networked approached where headquarters itself may be scattered throughout the network. Does it come as any surprise to any of us that the demand for larger systems is slowing while the demand for multiples of lesser-sized systems is on the increase? Should we be equally surprised if the proponents of this model are seeing more creativity?

There will be new challenges of course. Standards are crucial as is the gravitation towards commodity offerings. In talking with comForte mangers, the story I am getting from a vendor’s perspective is that the demand for uniform and secure data communications, better operational tools and services including one “pane of glass” visibility to all that’s in flight, as well the decomposition of applications into services for in-flight reassembly in support of new requirements, all become paramount considerations to better support the morphing of business that is occurring.

I am always cautious about reading too much into just one publication or pulling too much out of only a single survey. There’s so much these days that touches on “extremes” when it comes to technology. However, with the papers and research coming from the likes of the sources quoted here, and from what I am hearing from vendors remaining very close to enterprise customers – hardware and software, I have to believe a transformation is under way.

And if it’s not something of importance and relevance to what we are doing, then I can’t help wondering if our competition is already sensitive to the transition and are well down the path of convergence and transformation and producing more creative staff as a result!


  1. Interesting thoughts, and to me it seems that several sorts of transformation are underway. Today, NonStop is nowhere near the level of importance in overall IT that NonStop would actually deserve. So I'd believe that the number of large NonStop systems sold recently hasn't much to do with the overall direction that IT is evolving to.

    At the end of the last century, distributed processing was king and brought us - along with lower price tags for hardware - a number of great mix-and-match opportunities that directly led into overwhelming complexity. We ended up with scattered data and scattered processes, and with distributed server farms containing a hodgepodge of different software components that are pretty hard to integrate and to manage. Companies soon found out that in many cases, running many small IT islands based on cheap boxes is rather expensive and not very effective.

    Companies then went back to centralization. But they found out that just putting all those components into a single datacenter did not help much.

    Decision makers now get fed up with having to run complex server farms in their datacenters, and are calling for outsourcing and cloud computing. Simply put, they want somebody else to do the dirty work. If running those server farms was cheap, nice and easy, managers would certainly not want to get rid of these tasks - thereby reduce their own importance …

    The hey days of mix and match seem to be largely over, people have started to realize that complexity is the main source of unnecessary cost and risk burdening today’s mainstream IT. Major IT vendors like IBM and Oracle/Sun are building vertical hardware/software stacks to leverage that trend. IT is getting more proprietary again …

    IBM recently announced their new line of mainframes which provides not only the usual speedup and price/performance boost due to newer chip technology. Rather, the really interesting fact is a new hybrid architecture interfacing IBM’s AIX and Linux blades to the mainframe complex and centrally managing their workloads, thereby creating a kind of convergent infrastructure (and of course, a lock-in to IBM products).

    Oracle (who recently acquired Sun) also have a strategy of offering an integrated hardware/software stack to their customers, hoping to lure them away from mix-and-match implementations like Oracle databases on HP hardware …

    The Internet age asks for more reliability, availability, scalability and security than today’s mainstream server farms can deliver at reasonable cost. In this context, hardware price tags are less important - the picture would not change much if X86 servers were given away for free. Open source software is also for free, but it did not change the world completely …

    There are some car freaks who like to build their very own kind of car from many different components. Ordinary people who just want to go from A to B are better off with the car of their choice bought at the dealer of their choice.

    That’s the value proposition of integrated product stacks - which will become more popular in the future, after people will have realized that plain vanilla cloud computing isn’t the best answer in each and every case.

    Talking to a number of ISV’s these days, I have been convinced that NonStop fundamentals are still important – maybe more important than ever, in particular if you put IT security into the equation too. Provided that there is a common ground to start from (ie., Java and – where needed - suitable coverage of related functionality such as JBOSS) there is willingness to talk and willingness to test/deploy ISV products on NonStop.

    So now is the time to prepare (and do some homework !) for a larger share of NonStop in overall IT deployments, which will also lead to larger NonStop systems being deployed – either at customers or at service providers delivering special and very reliable cloud services based on NonStop (such as we are seeing in the payments industry since many years).

  2. Lot's to think of here, Gerhard - and up until recently this was exactly what I had been thinking. And maybe you have identified a key element - for most corporations, take it "off the shelf" or, as you suggest, take the integrated stack(s) on offer from metal to applications ....

    But I am not so sure if this (integrated stack, soup to nuts solution) plays to the creativity needs companies are looking for in order to innovate and separate themselves from the pack! In part this, too, is supported by my own assessment that NonStop will never be for the masses unless, of course, the NonStop platform transforms and becomes slightly more than a "controller" provided in every hybrid or megaplex as part of the “central nervous system” needed to ensure the other bits can be kept running, no matter what!

    I am getting a little ahead of myself, of course – however, last time I looked at the sky, there was never a big cloud, just lots of them and as I watched them, they would break apart, reform, disappear! And I have to wonder whether a well architected cloud as we think of for IT couldn’t indeed be broken apart and decentralized in a way that doesn’t compound the complexity while improving the flexibility and helping deliver the right amount of compute power / services needed to support the creativity I envisage will be needed to transform!

    What do you think?