Database optimization in the era of Big Data/Analytics and the Internet of Things (IoT) is a challenging task for data center managers. With data volumes growing rapidly, improving the performance of databases is becoming a key priority in data centers.
The Percona Live MySQL Conference and Expo in Santa Clara, CA, (April 13-16, 2015) showed IT developers in the MySQL and OpenStack spaces many ways to improve the performance of their databases in a multi-DB data center. Hybrid computing environments, linking enterprise and clouds – got much of the attention at the Percona conference, as this is the emerging business model for enterprise customers who retain their own data centers – but subscribe to cloud-based data services.
Just as clearly, security and availability of this data must be supported in modern data centers – whether the data resides in the enterprise site or in the cloud. After all, it is business-critical data for cloud-centric customers, including many of the cloud service providers attending the conference. Their infrastructure and their databases have the ability to accelerate delivery of their data services to end-customers.
The Tip of the Iceberg for Software-Defined Infrastructure (SDI)
Neuralytix believes that this is just the tip of the iceberg as the world of software-defined infrastructure emerges – bringing cloud-style computing to enterprise data centers, and introducing automation for end-to-end management of workloads.
It is already clear that scaling databases, and managing these large datastores is becoming difficult for many organizations, leading to a need for better analysis of the data already on-hand, and working effectively with it to gain key insights and to make businesses more efficient and agile, wherever possible.
Many of the 1,200 attendees came to the conference to learn more about “how-tos,” tips and techniques to work with open-source databases and OpenStack open-source software tools more effectively. Presentations by large service providers offered insight into ways to optimize MySQL and OpenStack deployents.
Presenters from high-profile web-services companies, included (in alphabetical order) Box Inc., Dropbox, Facebook, Google, Tumblr, Yahoo, and Yelp. Vendors were here, too, including Canonical, Cloudera, HP, IBM, Oracle, Red Hat, and VMware.
Keynote highlights included the following:
- Percona CEO Peter Zaitsev announcing the acquisition of Tokutek, a privately held company based in Lexington, MA, that supports highly dense data storage via its Fractal Tree indexing capabilities for MySQL and MongoDB. This technology improves data storage for both MySQL and MongoDB database management systems.
- Facebook presentation on “Polyglot Persistence” of multiple databases, by Harrison Fisk of Facebook, who said that variations in query size, data size and access patterns have resulted in deployment of a variety of databases for different Facebook workloads and data services.
- Facebook’s Steaphan Green delivered a separate keynote on WebScale SQL – the result of a collaborative open-source development effort between the MySQL teams at Alibaba, Facebook, Google, LinkedIn and Twitter. WebScale SQL adds more features to MySQL data repositories for scalable computing.
- Panel Discussion on “The Next Disruptive Technology: What Comes After the Cloud and Big Data?” with panelists from EMC, ObjectRocket/RackSpace, Percona, VMware and Yahoo. Key ideas for follow-on developments included leveraging data from the Internet of Things (IoT), using business analytics for data discovery and IT efficiency, and improved software management to make data centers more efficient.
- Discussion of new features for MySQL, by Tomas Ulin of Oracle. Sun Microsystems acquired the company MySQL, which developed open-source MySQL technology, in 2008; Sun was later acquired by Oracle in 2009. Today, Oracle adds new features to the source code of MySQL, then publishes them to the worldwide open-source MySQL community for production use and changes – with code fed back to that community, including Oracle. The current production version is MySQL 5.6, with Oracle making the release Candidate for MySQL 5.7 available to developers.
- Tesora founder and CTO Amrith Kumar spoke about building and optimizing Database as a Service (DBaaS) for the cloud. Tesora contributes to the Trove open-source project, which focuses on DBaaS.
- Apple co-founder Steve Wozniak spoke about the importance of innovation and creativity for problem-solving in application development, and in education.
- Discussion of VMware’s Continuent software by VMware senior director Robert Hodges (formerly of Continuent, which VMware acquired in October, 2014), explaining the ways it will enhance HA and DR for MySQL clusters running in the cloud.
Coping with Rapid Data Growth
The growth of data is becoming a central fact-of-life for data centers, with unstructured data joining the ranks of structured data for data ingestion, processing, data analytics and, ultimately archiving.
For many datacenters, this task threatens to overwhelm available resources, in terms of people, time and money. Technologies and best practices are emerging in response to this data tsunami, including those that replicate data, compress data, and manage it throughout its lifecycle from new transaction to aging cold-storage data.
At the Percona conference, presentations focused on:
- Optimizing MySQL performance, increasing SQL queries through a variety of technical approaches
- Ensuring data availability and security
- Linking on-premise databases with off-premise clouds and cloud services
- “Sharding” of large datastores, and workload-balancing across hundreds of servers supporting database processing and analysis
- Working with Hadoop data analytics
- Working with OpenStack open-source software, from hypervisors to middleware and databases, to applications, and management frameworks.
The top priorities for cloud computing and Big Data are ensuring the fast and effective delivery of data services. Those goals can be achieved through the use of hardware and software that, taken together, accelerate database performance. As such, it is central to Percona’s business of working with customers to optimize database performance.
Just as important are the traditional enterprise needs for availability, security, scalability and serviceability. They persist into the cloud era – and they are doubly meaningful for enterprises that combine their own IT data center infrastructure with data services from cloud providers.
These hybrid on-prem/off-prem deployments bring home the point that, while the technology specifics of computing have changed in the cloud computing era, the fundamentals of making sure that all data arrives safely and securely at its destination have never changed.