If you believe movie lore, a group of 300 Spartans held off millions of Persians at a bottleneck at Thermopylae. Historically, it is more likely that the deficit was not quite that wide. Either way, a business trying to account for all the data out there is sort of like a few Spartans trying to hold off an angry Xerxes-led attack.
This is perhaps what IBM was thinking with regard to Project Sparta, IBM’s simplified big data campaign. At a presentation in Singapore on Monday, Project Sparta turned into PureData, IBM’s latest offering in their Pure series.
PureData consists of three components: PureData System for Transactions, PureData System for Analytics, and PureData System for Operational Analytics. With the explosion of data in recent years (IBM agreed with Oracle’s assessment that 90% of the world’s data has been created over the last two years), IBM recognizes that one system cannot do it all.
IBM’s focus with these products is the simplicity and speed with which companies can implement these systems, as well as using their “built-in expertise” garnered from collecting and analyzing the ridiculous amount of data points they have. According to IBM’s website regarding their operational analytics database, “built-in operational analytics expertise, based on IBM’s years of experience and best practices from thousands of client engagements, is embedded into the system to provide a complete solution that can deliver value out-of-the-box.”
Of course, scalability is also paramount. All three systems claim to scale to at least a petabyte, with the most impressive being the transactional system. Per the website, “the system supports up to 30X scaling of cores and memory per database,” while also being able to consolidate up to 100 databases within the IBM platform. Further, the transactional database is available in three (small, medium, and large) iterations. According to IBM, however, if a company can upgrade from small to medium or large without any system downtime, providing in-system scalability.
Along with scalability, IBM stresses speed and simplicity with their analytics platforms. Their “built-in experience” comes in with regard to simplicity. So too does the limited amount of time it takes to receive the system and get it running some data. The transactional system is able to “deploy databases in minutes, not hours.”
The analytics platform, which runs on Netezza technology, reportedly offers responses “10-100 times faster than traditional custom systems. The system also, according to IBM, has built-in geospatial analytics, helping companies determine transactional locations. With regard to speed, the operational analytics platform is able to “handle 1000+ concurrent operational queries.”
These systems of course have several applications, especially in the healthcare industry (arguably one of the biggest sources of data growth and inefficiency) and fraud detection. “In the 200 or 300 milliseconds you have while the payment is going through,” said Arvind Krishna, GM of IBM’s Information Management, “you can do a quick fraud assessment.”