What is the best way to embark upon cloud computing? What do you do initial? We're carrying out a year-long contest on ReadWriteCloud, asking folks for their comments about their views about cloud computing. Every month, we'll give out a MacBook Air to the person who we believe supplies the very best comment on that month's subject.

There is nevertheless time for you to give your comment for this month's contest. Here's the theme:

A initial action towards creating a virtualized infrastructure typically arrives at a point when the buyer begins suffering from efficiency issues. Servers commence failing as software loads improve. When these problems begin taking place, what are the very first measures to take into consideration? How really should a virtualized infrastructure be deployed? How are these pre-production environments produced? What are the most successful techniques to deploy a simplified, dependable and optimized virtualization answer?

What do you think? Go away a comment on our submit for your likelihood to win a MacBook Air.

Here's a comment from Jose Gutierrez that supplies a excellent illustration for the sorts of responses we are hunting for:

We have not moved every thing to the cloud. By making use of the "do a single thing nicely, and then expand" technique, we found selected factors of running remote virtualized hardware that would have a noticeable damaging affect on our operations. The two downsides to the cloud that have so significantly kept us from relocating every thing are the network latency and huge information transfer times. We have been aware that the two of these could be issues, but by taking our migration one phase at a time (and starting up with people pieces that would not be affected by latency or huge information transfer instances), we ended up ready to determine how significant people concerns would be for us. Moreover, yet another downside to the cloud--comparatively slow I/O--ended up being a non-problem for us.

At this point, we have every single automated part of our information-processing pipeline on the cloud, but have kept the server that serves our neighborhood person interfaces and applications. So, in a sense, we have basically returned to a client-server setup, wherever the customers (and servers that handle customer interfaces) run locally for latency causes, and the servers that take care of heavy-lifting and require dynamic sources run on the cloud.

What do you feel? What is your contemplating about how to get started with cloud computing?