Once upon ago in the 1980s, before Microsoft Windows was invented, before IBM introduced the first PC, and more than a decade before Tim Berners-Lee introduced the World Wide Web, the first version of Oracle database Software went to market.
The version was called version 2 because Larry Ellison and partners feared no one would want to risk using the first version of a product.
Since that time, Oracle has sold its products and services to 20 of the 20 top airlines, 20 of the 20 top automotive companies, 20 of the 20 top banks, 20 of the 20 top governments, 20 of the 20 top high tech companies, 20 of the 20 top insurers, 20 of the 20 top manufacturers, 20 of the 20 top oil and gas companies, 20 of the 20 top pharmas, 20 of the 20 top retailers, 20 of the 20 top telcos, 20 of the 20 top utilities …
We could go on, but the point is made, Oracle, and particularly its database management offerings that took root more than three decades ago, rule many of today’s leading Enterprises.
That Was Then, This is Now
“That’s problematic,” says Robin Schumacher, Vice President of Product Management at DataStax, a company which sells and supports an enterprise version of Big Data database Apache Cassandra.
“Oracle wasn’t built for today’s modern, data-driven applications,” he says. Some companies who continue to use it for their strategic applications may lose not only market share, but also the shirts off of their backs.
Consider that at the time Oracle was designed, information wasn’t coming in at today’s high rate of speed; that it was coming into only one location, primarily from one location; that it was structured and well-organized; and that when volume grew too large, data was purged because storing it was expensive.
That is how it was yesterday; today it is no longer the case. Consider today’s rate of data velocity, the many different kinds of sensors that render data exhaust; the many different locations (including mobile) that data flows into and out from, the fact that the data doesn’t tend to be organized, and that global businesses need to be able to write and read everywhere at any time.
Traditional relational database management systems, like Oracle, can’t deal with the latter.
“Today’s Enterprises need to work with databases that are flexible, fluid and that can keep up with the pace of business,” says Schumacher.
And if a company’s database strategy is insufficient, the consequences can be quite costly. They range from losing competitive advantage to actually losing business; consider that Netflix went down for more than 48 hours when its Oracle database failed.
“We couldn’t risk that happening again,” says Christos Kalantzis, cloud database engineering manager, Netflix. “Oracle wasn’t built for the cloud and it doesn’t work in the cloud at the level we need it to,” he adds.
So what was Netflix, or many other companies who find themselves in similar positions, to do but to search for alternatives? Kalantzis says that his company considered databases that were built for a Big Data world -- MongoDB, Riak, HBase and Cassandra -- they also considered building their own sharded solution.
“Cassandra was our clear choice,” says Kalantzis. “It’s more elegant and easier to manage and scale,” he adds.
Today, Netflix stores 95 percent of its data in Cassandra, from customer account information to movie ratings, bookmarks, metadata and even logs. With more than 200TB of data in their production clusters, Netflix operates one of the largest cloud platforms in the world.
Has Oracle been totally eliminated at Netflix, we asked. Kalantzis says “no”; like many Big Data solutions, Cassandra has yet to be recognized by the government for being safe and secure. Certain financial data, according to Kalantzis, continues to be stored in Oracle.
Growing Pains in the Big Data World
Netflix is hardly the only company at which Datastax’s Apache Cassandra solution has replaced (or is replacing) Oracle. Quickly growing startup Ooyala, which powers multi-device video for the world's biggest media companies and brands, brought in Cassandra because it could predictably scale their massive and highly variable influx of data, which now reaches two billion data points each day.
Cassandra enables their database to run smoothly and scale without surprises because they can simply stand up a new cluster whenever they need more capacity.
“We quickly outgrew Oracle’s MySQL solution, and it became clear that relational database technologies were not an option because they could not support our analytics and broader big data initiatives,” said Sean Knapp, CTO and co-founder, Ooyala. “NoSQL big data solutions are becoming the default tools, and we chose DataStax because the Cassandra community was by far the most focused, had the most unified vision and demonstrated incredible execution.”
Why Go for an Open Source Solution, Anyway?
Sure, Open Source software is typically less expensive to obtain, (its basic versions are usually free) and use, but that’s not the only reason companies like it.
“Open Source tends to attract higher caliber engineers; code tends to be reviewed by others, and it’s a magnet for attracting the world’s finest talent,” says Kalantzis.
And when you’re pioneering a brand new field, the latter is a must; as is using software that was built for the age of Big Data rather than retrofitted.
Is sticking with legacy databases a business killer, we asked Shumaker. "If all you need to is run CRM and ERP applications in a single location, maybe not,” he says. "But as soon as you start looking at global, multi-location, mobile and high volume, you may have a problems," he adds. “And while you can try to shove a square peg in a round hole, why would you want to?” he asks. “Especially when you have a square peg at your disposal?”
Title image courtesy of Stephen Coburn (Shutterstock)