In the last couple of weeks, I was engaged in providing consulting to a company whether they should migrate some of their departmental functions to the cloud or not; if yes, how. The series of interactions I had with their team and recollection of similar incidents earlier with other organizations have incited me to think deeper into the methods and best practices involved in such migration. In this article, I would like to revisit the broad story as well as the unique parameters associated with organizations in their adventure into the Cloud (refer to the figure below for illustration). If you want to know about cloud computing, please refer to my other article, "The Story of the Digital Cloud - Cloud Computing".
It looks good when we use fancy terms like 'big data' and go on producing more and more data with the hope that these will be of some use in the coming days. But the reality is that we already have crucial data that are available in the form of office documents (created through MS Office or Open Office or other similar programs). We have digital assets in tape drives, DVDs, and of course, in hard drives, in the form of software files, videos, pictures, drawing files, etc. Above all, we have transactional data from relational databases that we have been patronizing for years (well, those digital systems have also served us sincerely and in great ways). And, we are tending to increase this pile unintentionally through our normal business operations.
In an organization, these data pose a challenge for the Administrator or Manager in controlling and auditing their state and transformation while data move inside the organization and inwards/ outwards. It is relevant when we envisage the scope of these data to fall in line with the business goals of the organization, and its usage requirements from staff, vendors, and customers. Moreover, the need for enforcing workflow guidelines and methods of collaboration widens the weakness of managing such data.
People want data available when they want to visualize the problem and make a decision soon after. If organizational data can be made available to the access device the user has, it can have tremendous utility. It is a ludicrous argument to brand the utilitarian goals of data usage to any recreational expectations. Then the demand is stringent. Make the data available in the right formats across the varying infrastructure of high and low bandwidth. You can't assume a cheap high bandwidth environment if your company sees a lucrative business in the foothills of Himalayas; uncertain KBPS connectivity is still the mainstay to the global information highway.
In short, your organization should classify the requirements along with the parameters of high availability, control/audit, and demands of workflow management and collaboration. But then there are a multitude of challenges if you want to implement your plan of adopting cloud computing as an extension of your legacy information system or a full migration to the cloud. Pragmatically full migration may be a myth as nobody would want to ignore the comfort of the rich interface of native Apps in the uncertain environment of disruptive connectivity to the cloud.
While discussing above, we basically intend to achieve seamless integration of data from the office desktop and servers to the central cloud that may be sitting somewhere on the globe. The integration demands adherence to certain I/O formats of data on the existing information systems or building such an interface. It's a task; and thus, there is a cost associated with it. Well, I am not going to discuss the details of the cost of cloud adoption in comparison to that of the existing legacy information system. In fact, there will be a competing argument of the recurring subscription cost, the re-migration cost in case of crisis arising in the cloud system, bandwidth cost at user-end, and above all, the cost of disasters. On the contrary, you may also argue about tangible benefits like waiver of upfront implementation cost (minus the migration cost as described above) and recurring maintenance cost. The arguments can go on. But then we must stop where we know the benefits surpass, making the way for overall profit. No aesthetics is involved here.
However, security concerns are paramount when we talk about the cloud. But then, why do we be worried about this perception if we rely on an ecosystem bound by strong SLAs (Service Level Agreements) and managed by specialized people adhering to a transparent set of quality and security processes. We should rather be happy by outsourcing security to a firm specializing in it. A win-win situation for both.
There are people from different organizations, especially government organizations, who would raise their eyebrows about shifting the data repository elsewhere. It is not their personal issue. Rather they are bound by certain compliance requirements. Such compliance factors negate the push for going into cloud. But in such a case, a cloud complying with their list of requirements will solve the issue. The government will want the data to reside within their boundaries; so will be the arrangement. Then we safely go with our decision about the cloud.
As I admitted at the beginning of this article, a decision will be guided by unique parameters that will influence the requirements, and by those that will challenge the implementation. Sad, but true, the parameters may vary; and more parameters may enter the scene with each individual organization in question. Well, that is why the organizations pay people like me; not a bad decision : )