Click to learn more about author Thomas LaRock.
The latest IT Trends Report revealed that 95 percent of North American organizations have migrated mission-critical applications and infrastructure to the Cloud over the past 12 months, with database workloads ranking in the top three. While the report found that many organizations are realizing the benefits of Cloud Computing, such cost efficiency, availability, and scalability, the findings also indicate that the resulting hybrid IT environments create new challenges, too. In fact, by weighted rank, the number one challenge created by hybrid IT is increased infrastructure complexity, followed by lack of control/visibility into the performance of Cloud-based applications and infrastructure.
Of all the applications and infrastructure being migrated to the Cloud, the database is of special interest. Applications are the heart of modern business, and databases are at the core of every application. As a result, migrating databases to the Cloud can often present more obstacles than expected. To help Database Administrators (DBAs) better prepare for and execute database migration to the Cloud, I will explore what DBAs most need to know, starting with the key considerations prior to migration.
Getting to the Cloud
With a variety of data and application parameters to consider, just getting to the Cloud is often a significant hurdle. There are two main paths to the Cloud: first, the classic “lift-and-shift.” This is where IT departments lift their current data and shift it to the Cloud, requiring little to no customization and architecting. The second path is a more involved process, requiring a DBA to start from scratch, building out and customizing infrastructure that will best meet the needs of its Cloud-hosted database.
Each path has its own challenges and benefits. For example, while the lift-and-shift scenario may sound very easy, many factors need to be considered to ensure application performance isn’t in jeopardy as a result of the transition. Similarly, without proper forethought and planning, DBAs run the risk of right-sizing servers incorrectly, or unexpectedly experiencing weakened performance because the database is no longer hosted locally—in fact, 35 percent of IT professionals reported moving applications back on-premises for this very reason.
However, this is not to say that starting with new, purpose-built infrastructure is a walk in the park. DBAs will still need to determine the best way to migrate their existing data and consider the risk of lost data. Furthermore, an organization’s Cloud Architecture is often designed outside the influence of DBAs, which is not ideal. DBAs should seek to take an active role early in the process, and if it’s too late for that, seek to understand why certain decisions were made and be a force for change if necessary. Doing so will help ensure that the Recovery Time Objective (RTO) and Recovery Point Objective (RPO) are suitable to keep the organization’s data protected. For example, they should help ensure geographical replications are in place to keep applications online in the event of an outage.
The relationship with the Cloud services provider and an in-depth understanding of the organization’s SLA with it are also important for DBAs to be successful in the hybrid IT era. Keep in mind that some Cloud services providers will accept shipped hard drives, some have dedicated high-speed connections, and some will even provide architectural guidance to act as your “database migration Sherpa.”
Once an IT department has successfully migrated a database to the Cloud, there are still challenges associated with managing and optimizing it. One of the hardest things for DBAs to adjust to is the lack of local access—which translates to control and governance—to the servers the database is hosted on, since they are now off-premises. This is a tradeoff, however, because with the loss of local access also comes (some) freedom from the daily management of the databases. However, the ultimate responsibility for performance and ROI still sits squarely on the shoulders of the DBA.
To that end, DBAs must understand all the various services offered by Cloud providers—while these services can aid in the management process and help achieve application performance optimization, the vast amount of services available to deploy means it’s possible an organization is paying for services that aren’t necessary.
Think of the plethora of HD TV channels available today. Just a few years ago, consumers paid specifically for fifteen or so HD channels. Now, HD is the standard and there are new channels popping up so frequently that more often than not, viewers are paying for access to them without even knowing it. It’s the same with the Cloud services.
Along the same lines, for any application or workload in the Cloud, it’s critical that DBAs hold providers accountable for the performance metrics outlined in the SLA. While Cloud Computing essentially means a workload is being managed by someone else, DBAs are still ultimately responsible for the success of Cloud-hosted workloads. As such, it’s key to regularly self-educate on new services and capabilities, review recommended architecture, and stay on top of scheduled maintenance that may impact application performance.
In short, it’s essential to think of the Cloud as a partnership—working with the Cloud services provider to ensure the needs of the organization are being met according to the SLA.
Best Practices for Database Cloud Management
Strategically moving a database to the Cloud can deliver great results not only for IT, but also the business’s bottom line. However, the challenges associated with getting there and then managing databases in the Cloud can be an intimidating obstacle to Cloud adoption. To help DBAs overcome these hurdles and experience a successful, optimized database in the Cloud, here are a few best practices to keep in mind:
- Strategy is Key: Before migrating anything to the Cloud—let alone a complex piece of infrastructure like the database—DBAs should take a step back and decide whether to lift-and-shift or create custom Cloud Architecture. There is no one-size-fits-all answer—it depends on the needs of each organization after making the determination which path will least negatively affect database performance. Leveraging a comprehensive monitoring solution that creates baseline performance metrics for all applications and workloads can help inform this decision. Furthermore, a roadmap that’s put in place prior to the transition will help cut workload costs and headaches. This includes having a fundamental understanding of the services provider’s SLAs and capabilities, as well as a thorough review of their recommended architecture and maintenance schedules. And remember, the Cloud is not an end-all, be-all solution. A database that is performing poorly on-premises will perform poorly in the Cloud if the root cause is not addressed.
- Establish Trust with the Cloud Services Provider: As mentioned, the Cloud should be viewed as a partnership where the internal IT department and the services provider are in sync to achieve the best results. However, at the end of the day, the DBA is still held accountable when performance metrics aren’t being met. Therefore, it’s important to “trust but verify,” implementing hybrid IT monitoring beyond what is typically offered by Cloud services providers. This ensures there is data and visibility to truly understand how workloads are performing in the Cloud and the reasons for that performance.
- Achieve a Single View: Today’s DBAs need comprehensive management and monitoring tools that provide a single dashboard of performance and the ability to drill down across database technologies and across deployment methods, including Cloud. This will ensure organizations aren’t wasting valuable budget chasing down a database performance problem with the wrong solution. With a broader set of data and greater visibility, Cloud DBAs can more quickly move through potential performance issues and correctly diagnose and resolve them.
- Be Mindful of Change: Software and services updates are now released at an unprecedented rate, with new bits often coming out every two or three months. DBAs need to think critically about these changes, as well as other trends in the IT landscape, constantly asking: “Is this something we can take advantage of? If so, is it the right choice or are there other options out there?” The ultimate objective has to be the best overall experience for the business and customers. This could mean it’s time for a change, but it may also mean the time is not right.
- Shift to a Proactive Mindset: Instead of reactively fighting fires, DBAs must transition to strategic roles designed to proactively improving the database. This will create extra time to gain knowledge about new technologies, while also cutting costs. Such a proactive, performance mindset requires looking beyond resource consumption and slow queries to performance-oriented methodologies like wait-time analysis.
Moving anything to the Cloud requires a shift in monitoring and management strategy, as well as paying attention to key considerations prior to migration. Databases in the Cloud in particular present DBAs several hurdles that must be overcome. By leveraging the above best practices, DBAs can find success in the Cloud.