

Last year Fusion-io introduced Atomic Writes, technology that reduces writes to only one process rather than two, and the company’s revealed their latest innovation – Extended Memory.
This subsystem is the result of a joint project with Princeton and is also available as a part of Fusion’s software development kit. The difference lies in what it does. It automatically tiers data from dynamic random-access memory or DRAM to ioMemory, creating a much more affordable alternative to storing that data in-memory.
Extended Memory automatically moves frequently accessed information to memory while transferring the rest to flash. This not only makes things more cost-efficient not just from the hardware standpoint but also on the developer’s end – a system only needs to be optimized for data that’s stored in-memory.
“The ability to optimize key operating system subsystems for flash with tools such as Extended Memory simplifies performance for developers in ways that were out of reach just a couple of years ago,” says Brent Compton, senior director of product management at Fusion-io. “We look forward to working with Princeton and other engineering leaders to continue simplifying complexity in application acceleration.”
There’s one other aspect to Extended Memory. Fusion-io says that using its SDK it’s possible to ‘lock’ data pages into DRAM, allowing the usage of flash as memory rather than an extension of disk storage. This can come in handy when scaling legacy apps.
Another recent flash story comes from SolidFire, which recently debuted a new, more powerful system. The SF6010 offers a maximum capacity of 2.4 petabytes and can scale up to 100 nodes. Each of these is an independent appliance of its own, making scaling a considerably less complicated ordeal than some alternatives.
THANK YOU