Ensuring Optimal Database Performance in the New Cloud World

Today, data management environments are highly complex and often span multiple vendors with deployments across on-premise data centers, clouds, and hybrid installations. In addition to the heterogeneity of systems, the processes surrounding database development and management have also changed. DevOps, a methodology for data scientists, developers, database administrators (DBAs) and others to participate in an Agile workflow, puts a premium on speed and also means that DBAs do not wield the firm control they did in the past.

Amidst those shifts, there is also greater need for stellar database performance due to the greater emphasis on real-time responses. As database environments are evolving, so is the job of DBAs. The movement to the cloud—along with more automation of tasks typically associated with the DBA role—has led to some speculation that organizations may need fewer DBAs, but, so far that does not appear to be true, according to a recent Unisphere research study, According to more than 60% of the respondents, the number of people with the title “DBA” is holding steady, while around 20% say the number of people with in that role is actually increasing. What is happening, however, with the greater diversity of platforms and methodologies is that data professionals’ lives are becoming more difficult.

Recently, Patrick O’Keeffe, Quest Software VP of Engineering, discussed how DBAs’ jobs are changing and what is needed to ensure optimal database
performance.

How has cloud changed the role of database administrators (DBAs)?

The cloud has changed the role of the DBA in a numberof ways. It has added a level of complexity. DBAs used to beresponsible for only a small number of servers on hardware.Then along came virtualized environments that took away hardware concerns, but added others. At the same time, the number of databases a single DBA needed to manage was growing. Cloud has just added to the responsibility mix—it is yet another environment in which DBAs have to managerisk as data stewards.

The cloud has also brought along new technologies for the DBA to master. Relational databases have been with us for some time, but the cloud has added database-as-a-service approaches for traditional RDBMSs as well as NoSQL offerings both in the transactional and analytics spaces. Lastly, the cloud has also ushered in new approaches to software development. Businesses want higher velocity at lower risk—this is the fundamental driver of DevOps cultures built on top of continued integration/continuous delivery [CI/CD] and Agile. The cloud both enables and tacitly mandates this culture—and configuration-as-code is de rigueur here. DBAs as data stewards have been slower to climb on board this train.

What are the issues with cloud that become more complex or easier?

The cloud takes away some issues and adds others. Take, for example, an organization adopting a platform-as-a-service offering. Things like OS patching and backups, to some degree, cease to be concerns, while data governance, security, and privacy start to become even more critical.

How is DevOps changing the DBA role?

DevOps is about putting creators and risk managers onthe same team, and this itself is new for DBAs from an organizational perspective. Businesses undergoing digital transformation are doing more software development and at the same time demanding higher velocity from these teams. This is demanding that DBAs work more closely with development teams.

This article is the third installment of a six-part series by editors of IOUG SELECT and DBTA on "DevOps and the Modern Enterprise" with three articles on the SELECT website and three articles on the DBTA and Big Data Quarterly websites. 

Is the fail fast mantra of the methodology a double-edged sword?

All innovation is change, which means that if you want to be innovative, you need to be prepared to change. You also need to be prepared to fail. Failure, always being an option, can be difficult for some organizations to confront culturally because often there are organizational or process constraints built specifically to try and avoid failure. So, for innovation to take place, the environment needs to be psychologically safe, and for that to happen these barriers need to be removed. This means that if you’re going to try to build a culture where failure is normal—and as it happens, necessary—you need to make failure as cheap as possible. Hence, the desire to fail fast. I’d say it’s absolutely positive.

What kinds of challenges does DevOps present for DBAs? Can you identify some of the dangers?

Tooling is an issue as databases are many, varied and complex. The tooling to support CI/CD for traditional databases is still not as mature as we’d like. This is potentially driving developers to use “simpler” databases for easier enablement of these workflows. This dynamic of development teams making technology adoption decisions is ascendant and represents business risk. The challenge is that development teams often do not consult experienced DBAs on these data storage adoption choices, and this is a missed opportunity because the initial choice can sometimes end up as technical debt that needs to be paid down at a future time.

What are some approaches that can be put in place to alleviate the risk associated with cloud and DevOps?

It is important to encourage DBAs as data stewards to become involved in DevOps initiatives. They have much to contribute due to their experience.
Having them involved early will mitigate risks around technology choice, and will ensure that requirements arising from needing to manage risk are at least on the table.

What are the capabilities that DBAs need to succeed in their roles today?

DBAs need to have an openmind, eagerness to learn, and the ability to adapt. Today’s DBAs should be ready to raise their hands and take a seat at the DevOps table. They should also be well-versed in all different types of databases—relational, non-relational, and cloud-based. As providers each
have completely different “SQL dialects” for programming within databases, DBAs also need to be fluent in each of them. Critical to the success of DBAs is staying on top of the latest tools available to make their jobs easier and create a positive impact in today’s data-driven business environment.

This article is the first installment of a six-part series by editors of IOUG SELECT and DBTA on "DevOps and the Modern Enterprise" with three articles on the SELECT website and three articles on the DBTA and Big Data Quarterly websites. 

To access article 1, go here.
To access article 2, go here.  



Newsletters

Subscribe to Big Data Quarterly E-Edition