Ravi Blogs about IT Topics

This question came up in the Manager’s meeting from Jolee. It is a good question and I began thinking more about this…

I wrote about Cloud Computing last October in this blog – click here to read it. I have become a fan of this concept and feel that this has a lot of potential if appropriately used. The answer to the question of whether “moving to the cloud” is outsourcing will depend a lot on the definition of each of these terms.

To me, both of them refer to identifying aspects of business processes that can be done elsewhere more efficiently and cost effectively. So far, Cloud Computing is more restrictive in that the typical services you get in the cloud are things like software as a service, and hardware cycles for compute heavy calculations. In most instances these services are provided as is. Outsourcing on the other hand provides a lot more flexibility and customization. However, the differences between the two are fuzzy and cloudy!

In terms of the the tests for efficiency and cost effectiveness – they should take into consideration all aspects of the processes and these tests by their very nature depend on the organization. We, as a higher education organization, cannot apply the same tests as a private corporation. For example, when it comes to IT related matters, the lack of standard computing environments and the importance of customer relationship building in our environment would dictate a different set of metrics and decision making than a corporation.

What follows are some random thoughts on this subject.

We already outsource some of our activities (or in some people’s eyes, “moved it to the cloud”) – for example, payroll and for recruitment. These are areas where our efficiency and cost effectiveness tests indicated that it is better to outsource it. Payroll is an extremely complex system with myriad of customizations and essentially zero tolerance for failure. Accurate paychecks need to be delivered on time to everyone, and we felt that in the overall scheme of things, this is better outsourced. Like all outsourcing, we have outsourced ONLY parts of the process – there are still staff members from HR and ITS who dedicate significant amount of time managing the payroll process.

Similarly, the Careers system helps us with recruitment and is hosted elsewhere. Building such a system in house and maintaining it would have cost us significantly more than what it costs to host it elsewhere. Again, all underlying information and the work-flow requires significant participation from everyone here including IT staff who are involved trying to integrate this system with ours to provide as seamless an experience as possible.

So, where does cloud computing fit into our future. I have some ideas, but I would like to hear from you (even if you think that we do not need it!). Some of the examples cited have to do with software evaluations. When we are ready to deploy a software, we usually scramble to find hardware to run it, test it and then size the systems. Almost always we end up using hardware that is already allocated to other services and this approach has resulted in problems several times in the past (though, with virtualization this will improve). On the other hand, the need to explore new systems all the time is a real one and is not going to go away soon.

So, why not do these evaluations on the cloud? Obviously it is easier said than done, because, many of the underlying infrastructure that we use (such as single sign on) would take a long time to implement elsewhere. But, if we structure it properly (by creating a smaller footprint of the most needed components on the cloud vendor’s system), I think this can work. It completely separates all the testing away from our production systems and it can be deployed fairly quickly without having to go through complex negotiations with others on the use of hardware.

I think the complex interconnected nature of our business (we need PeopleSoft and Active Directory for almost everything we do) would suggest that running any of our production systems on the cloud is not practical at this point. We should also thoroughly vet the data security and privacy before doing anything. In addition, we will need more network bandwidth. However, it is not inconceivable that standards emerge in the near future that enable secure communication between applications running on the cloud and some of the core “data” infrastructure on our network that will address these concerns adequately.

An even more futuristic thought is that our entire data center moves to the cloud. We certainly will be greener because we will presumably be sharing the compute cycles with many others. Obviously we will not even think about doing this unless our expectations for data security and systems availability are all guaranteed. But in a world like this, we will still have to do a lot of what we do now – we will be training the users, supporting, upgrading and maintaining various software and operating systems, writing applications etc. except, the hardware will be somewhere else and we will have a much fatter network pipe! And some of us will be spending our time on different things… Of course, first and foremost, this model needs to pass the efficiency and cost effectiveness test. We have a LONG way to go before we get there, if at all…

However, I think we should always look for opportunities in this space that will help us eventually enhance the core academic mission of the institution. In other words, if some aspects of what we do is better handled on the cloud, we should not be afraid to latch on to that opportunity and then devote our time to many other things that we can do to support the academic mission of Wesleyan.