Avoid these five cloud storage pitfalls
I’ll be honest; I’ve used cloud storage in some form for a long time. In fact, I started as early as 2009 with production data from my previous role in financial services being sent to a public storage cloud. Since that time, I’ve advanced a bit in what I think public cloud storage can be for me and what I do on a day to daily basis. In that process; I’ve learned a few things that I’ll share with you here.
#1 Lack of portability
The first thing I recommend that you avoid is lack of portability. What I mean by that is having a way to remove the storage that is in a cloud. There are many cloud storage gateway approaches that have a local cache and a copy in the cloud, which is common for files to the cloud and filesystem to the cloud gateways. Having a local cache can permit a co-existence in two clouds, or utter removal from a storage cloud, which is a great transition to the next topic.
#2 Getting locked in
You have to be able to change your mind on a cloud storage solution. This is something that we’ve had to deal with in on-premise storage solutions; the cloud should not be any different. Whatever cloud storage gateway, data mover solution, or management framework you are using for cloud storage should not lock you in to any specific cloud storage offering.
#3 Not enough bandwidth
The reality is that not all of us have the necessary bandwidth to fully embrace cloud storage. Be selective about what is a good use case for cloud storage. If you consider it for backups, make sure you have the bandwidth to return everything you’ll need in a timely fashion. That is part of my logic in a post I did last week, suggesting that there may be a cost savings with cloud storage, and the abstraction it provides can be priceless.
But if we don’t have the bandwidth to get everything back quickly, we may need to take a closer look. One way to do this is to leverage the rest of the cloud. Specifically, connections that may have very fast bandwidth to the cloud storage. That’s where my next point can help.
#4 Not taking advantage of compute resources
Cloud storage is one thing, and in the larger offering of cloud computing technologies compute resources in a public cloud may help you with storage situations. In most configurations, they have absolutely incredible bandwidth to the cloud storage resources. This means that if a large data profile needs to be traversed, it may make sense to do that in a compute resource in the public cloud. This can save a lot of time and possibly give you more functionality. Of course this depends on what exactly you are doing in the public cloud storage resource, but the takeaway is the cloud is there – use it.
#5 Forgetting encryption
With a few exceptions, encryption is a requirement for almost anyone who wants to venture into public cloud storage. Don’t go there unless you have a solution you can encrypt and manage the decryption process on your own. This will allow a lot of hurdles to be passed for data governance as well as your own peace of mind. My recommendation here is to take the time to understand encryption technologies available for your cloud storage application and apply your requirements to it. Then ensure that they work.