High Performance Storage Enables Digital Transformation

GridGain contributed to Apache Ignite and runs an In-Memory Computing Summit.  The company says that unified in-memory computing will power digital enterprise and bring on digital transformations.   In particular an in-memory store can assist with analytical and transactional data processing.  The Ignite integrated persistent store keeps all data on disk with all or some of the data in memory and processes done on data in-memory or on disk.  The fast in-memory data grid restart begins processing on disk data while the memory warms up and with the in-memory database data backed up to disk.

Gartner says that in 2019 70% of cloud-native application development is utilizing IMC or services that use IMC to implement high-scale/high-performance applications and that by 2021, at least 25% of large and global organization will adopt platforms combining multiple in-memory technologies to reduce their IMC infrastructure complexity.  They also predict that by 2022 90% of DBMS and data grid software will either support or depend upon non-volatile RAM (NVRAM).  Gartner market projections for the in-memory computing market are $18B projected for 2022 with a 20% CAGR.

ClearSky, a provider of on-demand primary storage with built-in offsite backup and disaster recovery (DR) as a service announced that the Myers-Briggs Company is using ClearSky’s service in the US and will soon deploy it for it’s UK officer.  As part of this plan, Myers-Briggs is migrating portions of its data center infrastructure into Equinix facilities.  

ClearSky and Equinix announced a partnership in November that creates a hybrid data fabric over the Equinix global platform, with ClearSky providing the data layer.   According to the ClearSky release, it:

·      Provides recovery point objectives (RPOs) and recovery time objectives (RTOs) measured in minutes; and

·      Reduces total cost of ownership for storage by at least 50%.

VAST Data announced version 2.0 of its Universal Storage Architecture (In March of this year we wrote about how VAST was using Intel’s Optane SSDs combined with QLC flash SSDs to provide high endurance storage system performance).  This product includes continuous snapshots, accelerated support for AI applications and asymmetric cluster expansion.  

According to the company, the key highlights of VAST 2.0 include: 

·Acceleration for AI Applications:VAST upgraded its NFS server to offer optional support for RDMA acceleration for high bandwidth machine learning and deep learning applications. With this new capability, customers running either Ethernet or Infiniband networks can achieve more than 20GB/s of single-client performance. VAST’s NFS offering delivers four times the performance compared to other scale-out NAS systems – ensuring that applications are never data starved and that expensive GPU investments are optimized for efficiency. The resulting solution provides the performance of a parallel file system, the IOPS and consistency of an all-flash experience and the deployment simplicity of a NAS appliance. AI, HPC and web-scale pipelines can now work faster, even in the face of rapidly changing application requirements.

·No-Compromise Continuous Snapshots: A modern approach to data protection that eliminates the copying of data or metadata during a snapshot event. Unlike legacy storage technologies, VAST’s snapshot engine is continuously running and allows users to instantly take snapshots with a very fine degree of granularity without compromising on performance, storage capacity or Flash drive wear – making it easy to protect data from accidental deletion or ransomware tampering.

·Asymmetric Cluster Expansion: Enables VAST’s Universal Storage clusters to expand online on an architecture that supports a diverse mix of hardware platforms participating in one unified namespace.