In-Memory Database Scales Massively For Social Networking Web Site
November 10, 2008 — Tagged, one of the largest and fastest-growing social networking sites, has adopted the eXtremeDB™-64 in-memory database management system (DBMS) from McObject® as cache for its commonly accessed Web data. In Tagged’s implementation, McObject’s 64-bit DBMS will host up to 100GB of user data in RAM to support features that help people meet and socialize, while also ensuring continuous operation and data consistency through its data persistence features.
On Tagged’s social network, eXtremeDB-64 will cache an in-memory copy of “all the connection information” used by members to meet, find out who is on-line, look up other people’s friends, and for other social networking functions. A comprehensive data set is stored permanently in an Oracle DBMS, but by keeping the most frequently used records accessible in memory with eXtremeDB, hits to the Oracle database are avoided and “we’re much, much faster,” said Johann Schleier-Smith, Tagged CTO.
Data held in memory is volatile, and Schleier-Smith said that a major reason for choosing eXtremeDB-64 was its transaction logging features, which add persistence to the in-memory records. Transaction logging writes database changes to a journal. In the event of a hardware or software failure, the eXtremeDB runtime can recover the database using this log.
A report published in September by Nielsen Online named Tagged among the top 10 up-and-coming social networking sites, serving more than 3.8 million unique users. Tagged ranked as the second-fastest growing, with 330% year-over-year growth.
The social networking logic that accommodates this expansion, using McObject’s technology, will reside on a middle tier of Tagged’s application, running on CentOS or Red Hat Enterprise Linux and multi-core Intel Xeon servers and managing approximately 100GB of user data across multiple eXtremeDB-64 database instances.
While in-memory database systems (IMDSs) were commercialized more than a decade ago, and caught on for embedded applications in part due to their typically modest memory footprint, recent benchmarks and deployments show the technology’s suitability for applications that demand both instant responsiveness and massive scalability. Examples of these applications include financial analysis and trading, telecommunications monitoring and credit and billing (BSS/OSS), and other solutions such medical and scientific research applications.
McObject introduced its 64-bit technology to provide in-memory management of databases hundreds of times larger than the 32-bit version of eXtremeDB. In a benchmark test performed last year, eXtremeDB-64 pushed past the known boundaries of in-memory database system (IMDS) scalability and performance with deployment of a 1.17 terabyte, 15.54 billion row database on a 160-core SGI® Altix® 4700 server running SUSE® Linux Enterprise Server version 9.
The benchmark demonstrated IMDSs’ breakthrough responsiveness as well as their ability to grow to a size that is well within the range of the largest on-disk Web and corporate On-Line Transaction Processing (OLTP) databases. Among other results, for a simple SELECT against the massive database, eXtremeDB-64 processed 87.78 million query transactions per second using its native application programming interface (API) and 28.14 million transactions per second using the ODBC SQL API. To put these results in perspective, consider that the lingua franca for discussing query performance of OLTP enterprise DBMSs is transactions per minute.
In the benchmark test, eXtremeDB-64 also exhibited nearly linear scalability performance from a single processor core up to 160 cores. McObject’s benchmark report, In-Memory Database Systems (IMDSs) Beyond the Terabyte Size Boundary, is available for free download and includes test results as well as query source code (using both SQL and the native eXtremeDB interface) and database schema from the benchmark application.
“Since McObject introduced eXtremeDB-64, its popularity has highlighted the gap in existing data management technology that can be filled by a highly optimized 64-bit in-memory database system. Real-time applications incorporating eXtremeDB-64 now range from commodity exchange trading platforms, to credit and billing management systems for vast mobile communications networks. We’re pleased that the product’s latest successful deployment will foster human connections on Tagged’s popular social networking site,” McObject CEO Steve Graves said.
Founded by embedded database and real-time systems experts, McObject offers proven data management technology that makes real-time systems smarter, more reliable and more cost-effective to develop and maintain. McObject counts among its customers industry leaders such as Chrysler, Lockheed Martin, Siemens, Phillips, EADS, JVC, Tyco Thermal Controls, F5 Networks, DIRECTV, CA, Motorola and Boeing. McObject, based in Issaquah, WA, is committed to providing innovative technology and first-rate services to customers and partners. The company can be reached at +1-425-888-8505, or visit www.mcobject.com.
McObject and eXtremeDB are trademarks or registered trademarks of McObject LLC. All other company or product names mentioned herein are trademarks or registered trademarks of their respective owners.