Get the highlights in your inbox every week.
Interview with Peter Zeitsev of Percona
The evolution of the database
This week in Santa Clara, Opensource.com is attending OpenStack Live, a two day conference with six hands-on tutorials and 18 sessions along with keynotes and panel discussions featuring a wide range of speakers including Apple co-founder Steve Wozniak.
OpenStack Live is co-hosted with the Percona Live MySQL conference, allowing attendees with interest in the cloud and databases to seemlessly learn from both events. So, I caught up with Percona co-founder and CEO Peter Zeitsev to learn a little more about the work which they do, how the open source database world and the world of cloud are converging, and what the two can learn from one another.
Percona has been working in the open source database space for a while. What inspired the focus on OpenStack?
I believe OpenStack has a chance to do for the cloud what Linux has done for server operating systems—basically becoming the defacto standard for designing modern applications. But it is not only the opportunity that is exciting: it is also about timing. OpenStack has great momentum with HP, Red Hat, IBM, Intel, Dell, and Cisco getting behind the OpenStack Foundation. The current state of OpenStack very much reminds me of the year 2001 for Linux when IBM announced it was investing a billion dollars in this technology.
Are there natural synergies between the OpenStack and MySQL communities? What might one learn from the other?
OpenStack is important for MySQL because it is going to be one of the most common platforms to run MySQL. The reverse is true also—MySQL is important for OpenStack as it is going to be one of the most popular DBMS running on it, and also because OpenStack itself uses MySQL as a backend store for many of its applications.
As OpenStack is moving from development environments to production environments, MySQL High Availability becomes increasingly important. At Percona we’re happy to be providers of one of the most popular solutions for this customer need.
Why do you think open source is important for enterprise database users and for OpenStack? Why is Percona dedicated to open source principles?
I think open source is very important for enterprises because it reduces cost and risk. Cost is obvious to most readers so I will elaborate on the risks a bit more. Proprietary software carries a lot of risks ranging from a vendor discontinuing the product, a vendor going out of business, or a dramatic price change that leaves users scrambling for other solutions. Security risks, especially having secret backdoors built in, is another worry for proprietary software. Open source does not eliminate this risk completely but dramatically reduces it. On the international scale, open source allows enterprises and governments to reduce another risk: the risk of being impacted by sanctions, when the proprietary software you count on suddenly becomes not available in your country anymore.
At Percona we make our core technology open source for exactly this reason—this is what smart customers now want and will continue to want in the future.
What do you think are the biggest challenges to sysadmins trying to manage databases in a cloud environment, and what is being done to address some of these challenges?
At this moment even how we’re running databases in the cloud is evolving. In the past we ran databases in the cloud, utilizing it as virtual machines controlled through APIs. Now, however, there are more and more databases being run as database as a service (DBaaS). For example, Amazon RDS is constantly growing in popularity and so is Trove in OpenStack.
The challenge for sysadmins is adapting to the DBaaS world which requires a different approach—you do not get as many controls and must manage things on a different level. Maturity is another problem—Amazon RDS is just reaching a level of maturity good enough for the majority and Trove is still pretty early on.
It seems like some of the fervor around NoSQL databases in recent years has diminished as developers are finding that they actually need many of the things that relational databases provide. Are developers coming back to the RDBMS, or did they never really leave in the first place?
I’m glad there is a change among developers going from the tools-first approach, where we pick the tool (Java, .NET, Oracle, or MySQL) and where we try to get the best pieces for the application we’re trying to build. We’re moving instead to a needs-first approach where we look at the overall requirements we have and then use the best tools for specific purposes. Modern architectures—SOA, microservices, etc.—allow the use of many of the tools at the same time successfully.
In some cases relational databases such as MySQL are a better choice than others like NoSQL or Big Data systems like MongoDB or Hadoop. In many serious applications, multiple technologies will be used side by side, known as "Polyglot Persistance." Facebook shared their story on that subject during a keynote at the Openstack / Percona Live MySQL Conference today.
I think we’re nearing the end of the hype cycle for NoSQL, and I expect a more thoughtful and balanced approach by developers when it comes to choosing technologies for their application moving forward.
This article is part of the Speaker Interview Series for OpenStack Live. OpenStack Live is a conference which is designed to teach attendees about the best practices and performance considerations for operating OpenStack, taking place in Santa Clara, California on April 13 and 14, 2015.