re: Expect To Save Millions In The Cloud? Prove It
Availability is an important factor as well. We use some cloud services to try them out and one fine day they out of a sudden were unavailable. Unfortunately, so was the backup plan. What happened? Big Joe with his steam shovel ripped the cables out of the ground and it took an agonizing long time to get service back. Since Big Joe did a neat job cutting fiber and traditional lines we could not fall back to a slower DSL. Now, what if we went all in with the cloud and ran business critical applications? Close the entire business for two days?
It showed that our disaster recovery was missing a big chunk. So now we mirror to a company owned data center a few states over and have almost all employees able to connect at least to shared resources and a virtual phone exchange. And we also keep only non-critical systems in the cloud. Who knows which disaster strikes on any part of the way between our office and where the cloud currently resides. Sure, the cloud itself may run fine and the vendors can switch over from a US data center to one in India or Sweden (legal issues aside), but if getting to it is a potential problem it becomes an issue. If constant availability is a must then the expenses for the plan B negate any savings from the cloud, both in cost and maintenance effort.
Another important factor is trust. Can we blindly trust the cloud vendor we chose? Are they really prepared to run our piece of the cloud 24x7x365? And are they trustworthy to handle our data properly? Someone sniffing around in it and stealing confidential info isn't even the top concern (encryption and adherence to policy can fix that), but what if Iwan the IT guy at the cloud vendor does a lousy job and our data gets lost or corrupted? Well, for that case we pull local backups which we hope to never have to use. While we have the data we still need the vendor to put it back where it belongs.
And then there is also cost. Yes, cloud potentially saves money, but some of the expenses are just shifted to a different place. Dragging data over local networks is fairly cheap and fast compared to pulling everything through external networks. We noticed a consistently higher data volume on our outside lines when people started using cloud services. That was not a surprise and we prepared for it, but there came the point were we had to upgrade the pipe increasing cost.
Time will tell if we really end up ahead compared to in-house solutions and it looks as if we do, but for a good part we just exchanged one set of problems with a different set of problems. And they are as annoying and frustrating at times to deal with as before.