It seems like the enterprise bugaboo of the decade is about to strike again. We've heard of server sprawl and then virtual sprawl -- now get ready for cloud sprawl.
But before I get too facetious, let me say that this is a potentially serious problem that can place a great deal of enterprise data at risk.
As described by Jamie Erbes, CTO of HP's Software and Solutions division, cloud sprawl is an outgrowth of what many enterprises are already experiencing with SaaS. With users across multiple divisions increasingly turning to the cloud for their computing needs, critical data could wind up lost once the project, or even the users themselves, are no longer active. She describes it as a shadow sourcing problem, which can best be addressed by updating policy procedures and deploying more cloud-aware management systems. Hopefully, both of these remedies will be forthcoming from HP.
Other organizations are starting to wake up to cloud sprawl as well. Over at Accenture Technology Labs, the focus is on making the internal architecture more like the external cloud, according to Director of Development Joseph F. Tobolski. In this way, management capabilities can extend across compute resources wherever they reside, while the enterprise maintains the flexibility and scalability that business conditions require.
Another player is newcomer Xangati, which recently released the AppMonitor suite that aims to improve virtual machine conflicts and sprawl issues without traditional agents or system probes. Although its approach is proprietary, the company claims it can resolve performance issues in half the time of traditional systems, and can even operate as virtualization and cloud services shift formerly static data architectures in new directions.
One of the reasons why cloud sprawl has stayed under the radar for so long is that much of the discussion has centered around "the cloud," according to The Channel Insider's Lawrence Walsh. But the fact is that there are likely to be many clouds, each of which will require unique browser sessions and credentials for various services. That means organizations could find themselves deluged with service contracts, SLAs and conflicting performance metrics.
Clearly, one way that cloud service and infrastructure providers could alleviate this problem is provide a means to purge their systems of inactive data on a regular basis, with adequate notification to the client that this is being done, of course. But even then, it will add another management burden to IT because they will be the ones to have to figure out what the data is and whether it is worth saving.
Ultimately, though, the cloud takes the problem of sprawl to a whole new level. Physical or virtual resources can be easily consolidated with the comfort that data is being preserved in-house. The cloud moves that data somewhere else, and it may not be quite so easy to get it back once the job is done.
