Cloud and outsourcing….

Today morning I was having a discussion with a faculty member on cloud computing – he wanted to look at some research in this area. The discussion moved to provisioning issues. Can we evolve better provisioning models that benefits consumers and vendors? While much of cloud computing is put under utility computing, there was some interesting connections that became apparent in my think aloud. Look at outsourcing cleaning services, catering services, computer maintenance, and so on. We can look at using a cloud in a very similar model. And a lot of lessons are shareable.

For example, if I need 3-4 full time cleaners all the time, then outsourcing is not all that worthwhile, and the outsourcing vendor also wont have much incentive, except for being a recruitment front-end for such staff. But if my requirements are scattered and a bit flexible, then with the same 3-4 people, the vendor can meet the needs of 2-3 institutions. If my daily (averaged) expense per cleaner is E1, I am willing to pay E1+delta per day, provided I dont pay for days I dont use the service. Since the vendor can put the cleaners to use elsewhere on those days, he manages E1+delta for almost every day, clearly earning some money. I, as a user, get more flexibility and spend less on the whole. A model that is a win-win for all the stake holders. This is exactly the idea that the cloud is based on, and has quite similar implications.

A further interesting dimension of this is the nature of measurement and granularity of deployment. In the case of cleaners, a day would be a good granularity. But, in theory, I can provide allocations on a per hour basis. The tradeoff is whether I can effectively use a resource that is freed for one or two hours. If you have a number of institutions within a building, then I can rotate people across places on an hourly basis, and hence shifting to an hourly allocation is worth exploring. The other part of the game, is the overhead in managing allocation at that granularity – I need mechanisms to monitor usage every hour, consolidate, and bill. If this overhead is manageable, and effective reuse is possible, you can reduce the granularity. This is one direction to improve provisioning in cloud – what to measure at what granularity.

Another idea is to look at the telecom companies calls being charged per second, instead of minute. Usually, this makes sense for under some assumptions only. And often the per-second charge will be higher than, the per-minute charge divided by 60. The rationale is that if you are used to making a lot of short calls (much less than a minute), then you pay only for those many seconds, and not a full minute. So you gain overall. But if you make a lot of long calls, then you end up paying more. So, this is an idea of introducing classes of users, and offer a deal that is more attractive to that class – knowing that users are in this class, allows the provider to position their resources more effectively.

There are more comparative threads between outsourcing in general life and outsourcing as practiced in cloud. It was an interesting think aloud, and thought of capturing some of those thoughts here. Yeah – nothing earth-shattering here.

Any thoughts from anyone?


3 Responses to “Cloud and outsourcing….”

  1. Srini Says:

    I like your analogy of cloud computing to cleaning services.. on the ground example! I am going to pass on the reference of this blog to folks in my office. 🙂 Thanks!

  2. padmini Says:

    i liked ur concept of cloud computing in cleaning services …….. nice example ..


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: