Predicting the near-future of IT
- 17 February, 2008 22:00
What's coming in "green" IT? There is growing recognition that IT devices consume an enormous amount of power. Technology vendors are beginning to make claims about their products' efficiency and power consumption, and yet there is no independent third party to certify these claims. But we will see the emergence of [such] bodies, and it will be a huge help to buyers.
What should IT managers do about the new social networking and collaboration tools invading the enterprise? We are seeing social media tools bleeding into the enterprise from the bottom up. These include wikis, document sharing tools, instant messaging and so on, and they are generally not sponsored by central IT. The early adopters tend to be project teams, particularly if the teams are geographically distributed.
But don't IT managers have good reason to resist the intrusion of these tools? They have valid concerns, but there is a lot of value in doing some experimentation to understand what the limitations and risks are and to use project teams as experimental testbeds.
Another opportunity that could have a big financial impact [is] for exception handling. In automated business processes, exceptions need to be handled by people — in extremely inefficient ways. For example, a customer has a non-standard shipping need. Often, these exceptions need to get a number of people involved. Who needs to be involved? Where are they? How do I reach them? Social networking tools, like a wiki, could be very helpful to resolve these issues.
What do you mean when you speak of the "consumerisation" of the tech industry? The old paradigm is reversed. A technology comes to the consumer space before it gets to the enterprise.
But again, don't IT managers resist that kind of thing? Yes, and that points to another shift we see, which is the increasing importance of business line managers as decision-makers about IT adoption. In many cases, the IT department is playing the same role that the corporate counsel plays: that of risk-watcher. [But] the business line managers are under extraordinary pressure for near-term performance, and they are saying, "I understand there are risks here, but I need this technology, and I'm going to put it in".
Might this shift of IT decision-making to non-IT managers signal the disappearance of the IT department entirely? Certainly, there are trends that will transform the IT department significantly. One is the trend toward outsourcing and utility computing. And as we move to much more modular, loosely coupled technology, it becomes feasible to pull apart significant parts of the technology and embed them in the business line functions.
But I want to caution against the disappearance of the IT department. There is still a huge role to be played in terms of evolving architectures to more effectively use technology. There is a healthy tension to be maintained between point implementations of technology to get near-term business impact and a perspective that says, How does this play over time, and are we getting huge inefficiencies if we don't pay attention to the architecture?
What trends do you see in outsourcing and offshoring? Offshoring is often seen as wage arbitrage, getting access to lower-wage employees. In many cases, quality suffers. But the real opportunity is in accessing world-class skills, even if at a slightly higher cost.
If you think of it as skills arbitrage, you need to [outsource] to multiple providers, because it is rare to find one provider with world-class skills in a full spectrum of activities.
What will be the next hot thing in IT architectures? We've evolved our architectures from the centralised glass house out to the desktop and, somewhat tentatively, out to connect to business partners. But what if you started with a clean sheet of paper and said, "I need an architecture to connect to 10,000 business partners, and I don't have the market power to enforce IT standards on all these partners"? What kind of architecture would I develop? There is a whole set of policies and management practices required.
[Service-oriented architectures] have tended to follow the traditional pattern from inside out, from a fine-grained transaction level. But increasingly, the challenge has been to maintain long-term relationships with distributed players and not just think about it at the transaction level. The IT architectures required to support relationships are quite different from those designed primarily to efficiently execute fine-grained transactions.
Do you see anything interesting coming in hardware in 2008? There has been a process that I see becoming more prominent, which is the componentisation of hardware — breaking down complex systems into much more commodity-like components. So rather than very complex servers, the trend in a lot of datacentres is to very modest, low-cost commodity servers brought together in massive quantities to deliver huge computing power. Companies like Google have been at the leading edge, creating datacentres with extraordinarily simple components to deliver extraordinarily complex functionality. There are real advantages in terms of flexibility and cost savings.