1. Higher than expected (and less predictable) operating costs
Lower costs are almost always touted as one of the biggest benefits of the cloud. But many organizations that migrate to the cloud are soon dismayed by “cloud sticker shock.”
Migrating to a cloud or hybrid-cloud environment promises cost savings in storage and maintenance, and is frequently a decision driver. However, cost savings aren’t realized unless resource usage is well managed.
For example, cloud providers often offer consumption-based plans, where the rate you pay is based on the bandwidth and services you use. Paying only for what you use is a no brainer, right?
It turns out that even though the cloud may offer more flexibility, predicting monthly charges is extremely difficult on a consumption-based plan.
A malformed query, for example, can end up costing thousands of dollars in unexpected and unplanned usage.
Overlooking the steps of optimizing queries and data scans can result in some expensive surprises.
Cloud providers also offer fixed pricing plans with a predetermined amount of usage available. But that same malformed query can eat up all your available resources if you’re on a fixed pricing plan. Just like going over your allotted minutes on a fixed monthly cell phone plan, you’ll have to pay additional fees for continued usage.
What you can do to make cloud operating costs more predictable
Pricing models will vary by cloud platform. Depending on the platform you’re using, you may be charged for the size of your database, the number of queries you run, the data being moved in a query, the number of rows in a query or a number of other variables. Therefore, before migrating your data to the cloud, you need to have a resource governance plan.
In some cases, you can manage your database directly. Snowflake, which offers a data warehouse built for the cloud, has a cool feature that creates separate data warehouses with different resourcing levels without duplicating data. This helps manage costs at a fine granularity, depending on the priority of the function or group. Caching technology and a central data governance platform is another way of managing costs.
In others, such as Google BigQuery, the amount you’re charged is proportional to the size of your database. But with an adaptive analytics fabric, you can build acceleration aggregates on your database, so you’re only charged for the data you use, not the size of the entire database. At AtScale, we’ve saved tens of millions of dollars for our customers by helping them avoid redundant full-table scans through our acceleration structure technology.