Friday, April 05, 2013

Cloud computing risk of data silo


Companies are more and more moving to the cloud. Cloud applications in the form of SaaS are getting more common in enterprises, hosting in the cloud in the form of DBaaS, PaaS or IaaS are getting more common by the day. I the"early" days of cloud computing companies who adopted where more small and medium businesses and startup companies. Now we see a move from enterprises towards the cloud. Large vendors are jumping on the cloud solutions and almost every day new cloud services and providers are seeing the light of day. The velocity of new companies getting created takes memories back to the days before the Internet bubble, maybe we are creating a cloud bubble at the moment.

However, we have seen that from a collapsing bubble also good things can come. When the Internet bubble collapsed the good where separated from the bad, this is something we might also see from a potential collapse of a potential cloud bubble.

Even though the cloud brings a lot of good there are some things to consider when moving to the cloud. I already discussed the legal implications of moving to the cloud in another blogpost. There are however more things to consider. One of the is related to a potential collapse of a cloud bubble. 

One of the things to consider next to the legal impact of cloud computing and the fact that your data might be hosted on a country that has a different legal system then the country where your company is located is the fact of data-silos. The risk of creating a data silo is on the agenda of a lot of the CIO's currently and it is indeed a potential risk of cloud computing.

When we talk about a data silo we talk about an cloud based application where we put in a lot of information however lacks the ability to extract this information. A large number of cloud based companies, and especially SaaS like companies do offer a great way of having your application in the cloud and provide you with options to enter information into this system. However the raw stored data is in most cases not accessible or in a very limited way. This makes that it is almost impossible to leave the cloud vendor and take your data with you to a new vendor. 

In cases where this is possible, based upon a contractual obligation or via build in functionality it often is a hard task to get the data in such a format that it is usable to be migrated to a new system. In cases where you have time to plan for a migration this is often costly however not impossible. In cases where this has to be done suddenly, for example due to a vendor that is bankrupt, you will not have the time to start a project for this. In those cases it might happen that you are unable to extract the data or extract it in a form that it is usable. 

It is good practice to check with your cloud vendor who is offering you a SaaS solution, or even a PaaS or IaaS solution, what your option are to access your data and what services are available to extract this. Next to this it is good practice to check with your vendor on what the rundown policy and exit clauses are in the contract in case the contract is ended and in case of a sudden bankruptcy. 

In an ideal situation your vendor has a escrow like mechanism in place for your data to ensure that your data is still available in case of a sudden bankruptcy. This will prevent you from data loss in such a situation, it will however not prevent you from service loss.

In case you want to extract your data during the contract period or at the end of a contract period it is ideal to have a mechanism in place for data extraction. Such a mechanism can be simple dump of all information in a pre-defined format, a direct coupling to the database holding your data or access to your data via an API.
Post a Comment