On-premises database-to-cloud migration is often seen as a silver bullet for cost reduction—really, how can you not save money when you don’t have to maintain all that infrastructure? Unfortunately, many IT professionals are learning that the cloud might not provide the level of cost savings they anticipated. There are still software licensing costs, varying charges for extra CPU and storage overages, and the common pitfall of over-provisioning the cloud service tier.
Make sure you ask the right questions
The new challenge is to strike the right cost-performance balance so that cloud computing actually does save more money than it costs. A good first step is to carefully consider the answers to these questions before you make a migration decision:
- Which cloud service is the best fit for your organization?
- Which of your databases is best suited for a migration to the cloud?
- How will you minimize downtime and avoid data loss during the migration?
- How will you monitor database performance and ensure scalability without blowing your budget?
- Do you have the tools to replicate data between on-premises infrastructure and the cloud?
- Perhaps most important, will you really lower your database licensing costs and avoid platform lock-in to any single cloud provider?
Fully understand your database workloads
No matter which cloud vendor you select, such as Amazon Web Services (AWS), Microsoft Azure or Oracle Cloud Infrastructure, you should be prepared to make a data-driven decision that involves a thorough understanding of your database workloads. Some of them might not run as well in the cloud and may not justify the cost of migration. Others may benefit from the rapid scalability of the cloud, such as for seasonal loads or offload reporting.
One way to obtain accurate data is to perform benchmark testing and performance monitoring on both real and simulated loads before migration. The results will allow you to right-size your cloud service tier and determine which databases will benefit from migration.
In this step, you assess the difference in performance between an on-premises database and one running in the cloud. You can use Benchmark Factory, part of Toad for Oracle DBA Edition and Developer Editions, to generate load and perform the same industry-standard test on both databases. Remember, the databases must be identical for a valid comparison.
In addition to a simulated load, it’s also good idea to monitor the performance of an actual on-premises database and a similar database running in the cloud. Here, the SQL Performance Investigator (SQL-PI) feature of Foglight can compare the performance of both databases and perform deep-dive analytics on where the database is being stressed. The simple, apples-to-apples comparison highlights any significant differences in performance between on-premises and cloud databases as user load increases.
Of course, you’ll want to continue to monitor database performance following the cloud migration. Here too, you can use Benchmark Factory to replay captured workloads to see whether running in the cloud has increased or decreased performance.
Foglight SQL-PI can be used to continuously monitor database performance so you can spot any potential problems with cost, performance or scalability in real time. Running real application transactions in both locations will help you see whether you have chosen the correct cloud service tier—the combination of compute, memory, storage and network—for the performance you expect.
Migrating databases to the cloud isn’t a silver bullet. Beside the operational risks of downtime and data loss is the financial risk of a mismatch between performance and cost. That is why Quest offers a complement of tools that provide the data you need so to determine whether the cloud is really saving you money.
To learn more about the tools that can help your cloud migration go smoothly, read our latest tech brief, Balance Cost and Performance when Migrating On-premises Oracle Databases to the Cloud.
Share this blog with your peers
If you think your colleagues would benefit from this blog, share it now on social media with the buttons located at the bottom of this blog post. Thanks!