#memeconnect #emc #bd
The new IT appliances such as Oracle's Exadata and the big data phenomenon are coincidental revolutionary trends in IT that only intersect incidentally. However, they do have one important thing in common. They both challenge the traditional siloed IT organizational structure, and they reinforce the pressure on the organization generated by the need to treat applications holistically to provide better service to business end-users.
IT appliances such as Exadata have several important advantages, including their much lower cost. But they create an organizational headache. Because they are essentially “IT-in-a-box” containing processing, storage, network connectivity, operating system, and applications in a single package, they do not fit into the traditional IT organization. Early adapters Wikibon has talked to have tended to assign their Exadata boxes to the storage group, but that is really a forced fit because the box is not only or even basically a storage device. It is an everything device, closer akin to a PC in that sense than to traditional IT components. Assigning it to storage forces the data gurus to deal with application, network, and other issues outside their expertise, but the same would be true of any other IT group.
What makes more sense is to build a multifunctional team to support whatever application is running on the appliance. This team should include skill sets covering all the basic IT infrastructure components plus the application itself. This cross-functional team would be dedicated to providing specific functionality to the end-user at a specific service level.
Some visionaries have advocated this kind of organization for several years to solve the problems of providing better service. The problem is a basic disconnect between IT professionals, who measure performance in terms of server uptime, network transmission speeds, disk seek times, etc., and business end-users who see IT in terms of specific functionality that they need to do their jobs. The argument is that IT organizations need to reorganize around the applications and functionality sets that they provide and see those holistically rather than focusing on the different speeds and feeds of the components. When all those components are in one physical box, the idea of reorganizing into cross-functional teams around specific applications or end-user functionality becomes even more logical.
Big data may have an even greater impact on the IT organization in the end for two reasons. First, big data projects are almost always stand-alone efforts, largely isolated from normal IT. They often rely heavily on third-party, sometimes publicly available data sets and other resources. And because they are often experimental, short-term, and high-risk, and the next big data project may require very different resources than the present one, it often may make sense to rent cloud resources rather than building a system in-house. And they are usually completely driven by business need and a business executive champion rather than by the CIO. They also are heavily analytic in nature – the idea is not just to capture and store data but to create business value through advanced data analysis. This means that they often will require teams that include data “quants”, analytics geeks, cloud service providers, and business people, none of whom are part of the IT organization. Ironically the IT team supporting these efforts is once again likely to be multi-skilled. A siloed organization will be unable to provide what these projects require.
Action Item: CIOs should start experimenting with cross-functional teams to support specific applications or end-user groups, particularly if they are also bringing in IT appliances, becoming involved in big data projects, or simply feel pressure to improve the quality of service to end-user groups. By looking at applications holistically IT can better understand the end-user view of computing.
Footnotes: