A client of mine recently faced the twin challenges of replacing its aging ERP system and improving its content-based business processes to accommodate its recent explosive growth. Everyone had opinions on what kinds of software to acquire and where to host it. But one of the other critical subjects was how best to migrate its existing storehouses of information, both paper and electronic, without unduly burdening its existing resources....
And so the question arose:
Is there a role the cloud can play in enterprise data migration?
Plenty of parking
Whether my client chooses to locate its servers on-premises or in the cloud, the data is going to have to move. Before that, it really should be cleaned. But if the current systems are already creaking and clogging, where can that clean data be parked while the new solution is being selected and readied?
One answer is to store that data in the cloud. To keep the operational data sets from becoming out of sync, work on the data in logical chunks. That way, the data can be continuously migrated to the cloud, bit by bit. This adds the complication of modifying the current systems to seek that data from its new server, so as not to work with the wrong information. But parking the good stuff means it can be safely separated out and managed on somebody else's infrastructure until the company is ready to integrate it into the new solution.
Divide and conquer
Sometimes the catalyst for migration is more directly related to resource optimization than data cleansing. One IT executive I spoke with described this in terms of splitting data into "the stuff we use all the time" and "the stuff we have to keep but don't want to crawl through all the time." Now, which gets stored where depends upon response-time requirements, Web-accessibility needs and general mindset toward the cloud. But the approach can work either way, and is really just a philosophical extension of using SharePoint or other in-house solution to help take a load off.
This strategy of division can work well when moving a system to the cloud rather than using the cloud to support an on-prem migration. In this case, splitting the load can mean putting a subset of data in one or more cloud solutions. That way, the company doesn't have to rely on somebody else's idea of what a useful demo is, in order to gauge their suitability. Just be sure the test data is representative of the company's actual data, but representative enough to divulge sensitive information. After all, outside eyes will likely be on it while the system is put through its paces.
The primacy of process
Anyone who's ever been close to an enterprise data migration project knows well just how many moving parts there are. Adding the element of cloud only complicates matters because of the myriad other considerations it involves. To help cut through the fog, consider the nature of the business problem that needs to be solved, and work outward from there.
My client's technological journey may involve an old ERP system, but it began with a business imperative: the need to have an information infrastructure that supports its market success. Recognizing this, the company hired an analyst to examine its business processes and what information is utilized by whom, at what time and under what circumstances. The result was a highly effective roadmap that is now being used as the fixed point around which all discussions revolve.
It is not yet clear what the organization's new architecture is going to look like, but having taken the time to really understand what its employees need in order to do their jobs, there's little doubt that the inevitable migration to come will go more smoothly than would have otherwise been the case. And if I had to guess, I'd say that one of the cloud-inclusive techniques touched on above will be in the picture.