As the scale of OLTP transactions skyrockets, disk-based databases are unable to keep up with demand. Here’s a look at the problem and how in-memory databases can address it. A new generation of in-memory databases takes advantage of modern, scalable multi-core CPUs to deliver OLTP performance in a way no disk-based database can.
Coming off of last month’s announcement at Structure of a new customizable chip loaded with a FPGA, Intel has shipped off a version of its Xeon E7 x2 processor that links together Oracle’s software to its hardware.
MemSQL, the database startup from two former Facebook engineers, has already raised a lot of money and roped in some big customers. Now it’s looking to broaden its footprint with a flash-optimized columnar store to complement its in-memory row-based one.
Yale researchers Daniel Abadi and Alexander Thomson think they have developed the cure for Oracle and IBM dominance in the world of database performance, and it isn’t even technically a database. The two have created a system they think can level the playing field.
According to database pioneer Michael Stonebraker, Facebook is operating a huge, complex MySQL implementation equivalent to “a fate worse than death.” It’s actually a predicament all too common among web startups, for which the solution might be a class of databases referred to as NewSQL.