Whether your company is born-in-the-cloud or you have a corporate initiative to migrate more data to the cloud, there are a few risks and challenges with putting your “crown jewels” into the hands of AWS, Azure, or GCP. In this post, we’ll outline what to look out for and how to circumnavigate around these challenges.
The Performance You Get
If you’ve previously managed your databases on-prem, this won’t come as a surprise to you: the performance that you are able to achieve in the cloud does not even come close to the high level of performance that you can get on-prem. In fact, if you want to try to match your on-prem performance levels, you will pay dearly to get it. In order to get the performance that you need, you will need to overprovision and pay for capacity that you don’t want.
But what if paying for extra performance (and resources) is out of your budget? Then you’ll end up compromising for lower performance which can result in queries timing out, batch jobs overrunning, and your end users getting frustrated by a sluggish experience.
Achieving better performance may sound like a problem that you can just throw money at to solve. But for more intensive workloads, you won’t find any suitable solution in the cloud for achieving the ultra-high level of performance that is required.
Having access to your databases whenever and wherever you need it is critical. But time and time again, we read in the headlines that one or the other major cloud vendors has had an outage. This most recently happened when AWS went down the day before Thanksgiving, the unofficial start of the holiday shopping season in the US, leaving their retail and ecommerce customers scrambling. Moving your databases to the cloud means you are risking that you’ll lose access temporarily and, Murphy’s law, probably when you need it the most.
If you want to create clones, snapshots, or copies of your data in the cloud, you’ll quickly find that you’re being nickel and dimed. Since each clone and copy takes up capacity in the cloud, you will need to pay for that.
Meanwhile, licensed database software – like Oracle or Microsoft SQL Server – set the license cost proportional to the number of CPUs being used. In the public cloud, you need to overprovision your VMs with more CPUs to get the performance that these databases demand. There goes a few more dollars down the drain
Migrating (To the Cloud and Back Again)
If you’re looking to migrate your database from on-prem to the cloud, the migration process can be a headache. As part of the migration, you might need to refactor your databases and applications to handle the new cloud architecture and the different performance challenges. But this process can be time-consuming, costly, and potentially risky.
And if your database is already in the cloud and you are looking to migrate it across your hybrid or multi-cloud, the same issue arises. Once you’ve moved your database to one cloud, it’s a struggle to move it anywhere else.
Overcoming These Challenges
The Silk Platform is perfectly poised to help you overcome all of these challenges. How? Silk is a performance supercharger, giving your databases significantly greater performance than they are able to achieve with native cloud alone. In fact, in performance tests, we have seen that MS SQL on Silk can achieve 5x higher throughput and IOPS versus native cloud alone, while ElasticSearch can achieve 3x higher throughput and 2x faster performance on Silk versus native cloud.
Silk helps keep your costs to a minimum through the use of zero-footprint clones and by thin provisioning your data. And in the case of licensed databases, Silk minimizes the number of CPUs you need so you don’t need to overprovision for performance. This offloads the compression from the VMs, thus decreasing the number of CPUs required.
Silk also offers 99.9999 resiliency and no single point of failure which gives you the confidence that your databases are always readily available. In fact, since Silk lives between your cloud infrastructure and your database – virtually keeping your databases on the cloud without entangling the two — if one of the cloud vendors has a problem in one or more of its regions, you can quickly and easily run your production from another vendor.
So why underperform when you can Silk instead? Learn more about what the Silk Platform can do for your databases in the cloud by visiting www.silk.us.