You don’t have to convince Jason Atlas that yesterday’s databases weren’t built for today’s world. The Vice President of Engineering at IID experienced it for himself when the popular database his company was using almost puked because it couldn’t handle the momentum of data coming at it.
And while a little latency might be tolerated in some industries, at Atlas’s company it is not; after all it provides a platform where enterprises can share data about the latest cyber threats and ongoing attacks and be warned before the trouble that’s on its way -- and it’s always on its way -- can wreak havoc.
Doing this requires gathering and crunching huge amounts of data from a large number of sources, so much so that the term “big data” and its three v’s (high volume, high velocity, high variety) seem barely sufficient to describe it. And this data, mind you, not only needs to be collected, crunched and analyzed, but this also needs to be done in a highly secure and transparent way so that the financial institutions and other businesses, who share their information with IID, have full confidence that it is protected, uncompromised and used for intended purposes only.
The stakes here are extremely high, according to Atlas, because it’s only by sharing data about attempted intrusions, threat indicators and confidences that IID’s customers stand a chance of shielding themselves against trouble.
The criminals are as organized and probably more capable than we (enterprises) are. They’re working together and we’re not,” says Atlas. “They share with each other in ways that we don’t. There are sites out there where criminals buy, sell and trade things like bots, worms and Trojans.”
Suffice it to say that IID, which sometimes acts as “mission control,” needs to leverage the best technologies the world has to offer to get its job done. Speed, 100 percent up-time, security and transparency are the basic requirements and traditional databases like MySQL and Oracle won’t suffice because they weren’t built for this data driven world.
So when Atlas looked to replace the database he was using, he considered a number of alternatives -- MPP databases and most NoSQL databases, but they didn’t meet his needs until he found one that did -- open source Apache Cassandra.
“NoSQL does not equal NoSQL,” he explains, noting that different NoSQL databases are built for different things. One is not necessarily like the other and they don’t always compete.
Atlas chose Cassandra for IID because he was sure it could get the job done. It offers 100 percent up-time, scalability, transparency and high availability, without compromising performance, as well as proven fault-tolerance on commodity hardware. For mission critical data, like he works with, it was a no-brainer, says Atlas.
But there were other things on Atlas’s nice-to-have list that open source technologies don’t come with when you take them “off the shelf” (and mind you, the Apache open source version of Cassandra is free), like enterprise-grade support, training, integration with other platforms and technologies, and so on.
So Atlas, who says he’s be terrified of working on mission-critical projects without commercial-level support, had to find a remedy. The salve? DataStax’s Enterprise platform.
Datastax, the most popular commercial vendor that supports Cassandra, fills in the gaps between what Cassandra has to offer and what enterprises need with its Enterprise edition. It provides not only an enterprise-grade version of Cassandra, but also enterprise grade Apache Solr for searches, Hadoop for analytics, and an OpsCenter which Atlas says is worth its (metaphorical) weight in gold.
DataStax Adds In-Memory Computing
Now that big data is becoming the norm, the next buzz in the world of computing will be, and, perhaps already is, NoSQL, referring to the “Not only SQL” databases that will empower the data driven enterprise. And while this new breed of databases are already powerful and fast, it’s only natural that we want them to be even faster.
After all, latency is expensive.
DataStax founder Matt Pfeil drives the point home when he explains that for every half second Google takes to return a result, it loses 20 percent of its traffic. If Amazon slows down by a tenth of a second, it loses sales.
So when you’re doing business in a world where time literally costs money, speed is a differentiator. And needless to say, the tech vendor who enables its customers to do the most business, in a smart way, will inevitably win.
DataStax Brings Speed to the Data-Driven World
DataStax is now bringing in-memory computing to its customers who use DataStax Enterprise 4.0. It’s an option (you can choose to use it or not, it doesn’t cost anything extra) that allows administrators to decide whether to assign data to in-memory objects, spinning disks or SSDs all in the same database cluster, making performance optimization simple.
This is something that e-commerce, telecom, financial institutions and cyber security companies will no doubt want to leverage. In some tests, the in memory option was as much as 100 times faster than Cassandra, according to Weist.
And that matters.
“Sometimes being slow is just as bad as being offline,” says Weist.
When DataStax Does What’s Best for its Customers, Oracle Loses
And while Datastax is the first NoSQL vendor to offer an in-memory option, it’s not taunting its NoSQL brothers with the news.
That’s because the guy to “beat on the street” isn’t another NoSQL vendor, it’s Oracle, who tries to keep beating its old dog into running faster.
Four-year-old DataStax now counts more than 20 of the Fortune 100 as its customers, so there’s no need to ask Larry Ellison how that’s working for him. And who knows, maybe he’s not thinking about it anyway. It seems that he’s been busy taking over the Hawaiian Island of Lanai, he now owns 98 percent of it.
But that’s an aside, of course.
Datastax Adds Search and OpsCenter Enhancements
Datastax has also made enhancements to its Search product and OpsCenter feature which make them smarter and faster. This will no doubt delight developers and admins.
The Right Tools for a Data Driven World
While it took a great deal of genius and innovative thinking to create the databases that have driven the world for the last 25 years, new tools are needed for a data driven world. Cool as NoSQL and in-memory databases may be from an engineer’s point-of-view, they are a “must have” for business. Sure, the fact they run on commodity hardware and are, for the most part, open source, represent big wins, but there’s something else to consider: 10 years from now you may not be able to do business without them.