ShareBlog

I have run into this a lot both as a consultant and just talking to people. People think that just by moving their sites, VMs, or entire company into the cloud then everything will magically take care of itself. The funny thing is I have heard this before with other “magical” technologies and it never happens.

Now anyone who knows me knows I am a HUGE proponent of cloud technologies (particularly Azure) and I think the cloud is going to become more and more important as time goes on. That does not mean that I think everything should go to the cloud and that just by throwing VMs and sites into the cloud it will automatically make them work great. Yes, the cloud has some amazing capabilities around scaling, disaster recovery, and other features but they need to be planned out.

I was talking to one person who was telling me how much the cloud sucked and how they pulled their sites back off the cloud because they were not working correctly. After much discussion it turned out that they had chosen a very small application service plan and never setup any sort of scaling so they were stuck with the small application service plan. While it worked great a lot of the time, the site just crawled when a lot of people were using it, as expected.

So should this site have gone to the cloud in the first place? I cannot say with 100% certainty that it should but based on the discussion I had I believe it was a great candidate for the cloud. The problem was the person 1) didn’t educate themselves on how the cloud really works and 2) did not make any sort of plans on how to move the site and what constitutes success.

I even had one person ask me how to connect a input device to their cloud VM. Well, you can’t. First, you have no idea what machine your VM is actually running on and it could actually change from one day to the next. Second, very few people have access to the Azure data centers and even less have access to the actual physical servers. Third, assuming points 1 and 2 did not exist how did you expect the machine to work? How would you get the information that needs to be inputted into the machine which could be thousands of miles away.

Now this is an extreme case and I seriously doubt I would run into it again but it does show that not everything can be moved to the cloud. Machines that require direct connections to external devices is a great example. Others include data that is deemed too sensitive to go into the cloud (I will not get on my soapbox about cloud security versus on-premises security but lets just say I do not like this reason but it is still a concern for a lot of people), costs (if you just purchased new hardware does it make sense to idle it and use the cloud? Probably not), old operating systems (not sure any cloud provider can support a VMS box), and others.

That brings me to my final point for this post. Why do you want to go to the cloud? Is it because it is the buzzword of the day? That is usually not a great idea. Is it because you heard it can save you money? OK, that is a better reason but still needs research to validate Do you have an overworked IT staff and want to take some load off of them? I have heard this once before, and was a bit shocked when I heard it :), but again while this is a valid reason, research needs to be done to make sure this will work. I think you will find that in the short term there may be more work while settings are being tweaked but long term the workload will decrease.

Just remember, the cloud is a great tool to help you achieve your goals but it will not solve every problem.