THE SOURCE FOR TECH BUYING ADVICE
SK hynix has something fun in store for attendees at the Flash Memory Summit Semiconductor company SK hynix is preparing to demo a new type of storage device, developed in partnership with the US Los Alamos National Laboratory (LANL).
The business plans to unveil the first Key Value Store Computational Storage Device (KV-CSD) in the world at the next Flash Memory Summit. This device’s goal is to speed up the processing of massive amounts of data.
By integrating indexing capabilities into the storage device itself, this goal is accomplished, allowing data to be written and indexed virtually instantaneously. According to LANL, the method can up to 1000x accelerate some of its simulation data analysis tasks.
Computational storage with a twist
Since it reduces bottlenecks caused by the need to transfer data between storage and the CPU, GPU, and RAM, computational storage has been hailed by industry players as one of the upcoming big things in computing.
These systems can be broadly divided into two categories: those that integrate processors within the storage device itself and those that send compute activities to a storage accelerator nearby.
Although computational storage is not suitable for all use cases, it has the potential to significantly speed up systems whose I/O performance rather than their compute performance is the bottleneck.
“There is clearly a broad class of applications that benefit from offloading compute functions from a main CPU to a more efficient processing engine that is more suited to the specific problem of interest,” said Richard New, VP of Research at Western Digital, in conversation with Epblogs last month.
“In the context of storage, we can think of applications like video transcoding, compression, database acceleration as falling into this category.”
Subscribe to Our Epblogs Telegram Channel instant breaking news, deals, opinion to stay up to date on the latest news and reviews.-epblogs
Despite the fact that several businesses are working on developing computational storage devices (CSDs), this is reportedly the first instance where indexing functionality has been successfully included into a CSD.
The creation of the KV-CSD is a reason for celebration for all enterprises handling large-scale simulation data and involved in big data analytics, according to representatives of SK hynix and LANL.
According to Gary Grider, who oversees HPC at LANL, “demonstrations like this demonstrate it is possible to build an ordered KV-CSD that moves the ordering and indexing of data as close to the storage device as possible, optimizing the wins on retrieval from on-the-fly indexing as data is written to the storage.”
“The ordering feature permits both the point searches that key value storage is known for as well as range queries that are particularly valuable in computational science applications.”
For the time being, the KV-CSD is still only a prototype; more development must be done before it can be used at LANL in a real-world setting. Also unknown is whether the conditions of the collaboration agreement could prevent SK Hynix from releasing the technology on a wide scale.