Part of IBM’s Software Defined Storage solutions, code named Elastic Storage Version 4.1, is now released ‒ delivering flash accelerated, software defined storage for cloud, big data & analytics, and technical computing ‒ now with enhanced security, performance and usability. Elastic Storage is built on IBM General Parallel File System (GPFS) technology.
New advanced features in Elastic Storage 4.1:
Native encryption and secure erase: Elastic Storage 4.1 introduces the protection of data at rest and secure deletion. Elastic Storage encryption is designed to provide protection of data from security breaches, unauthorized access, and being lost, stolen or improperly discarded. Cryptographic erase provides fast, simple and secure file deletion. Elastic Storage 4.1 complies with NIST SP 800-131A and FIPS 140-2.
Elastic Storage Flash Local Read Only Cache: Elastic Storage flash local read only cache accelerates I/O performance up to 6x by reducing the time CPUs spend waiting for data and reducing the overall load on network and storage resources. This improves application performance while still maintaining all the manageability benefits of shared storage. Elastic Storage handles the flash-based cache automatically so data is transparently available to your application with very low latency ‒ more on this exciting new feature later in the blog.
Network performance monitoring: Optimal Elastic Storage performance depends on proper network performance and behavior. Elastic Storage 4.1 now provides for monitoring of network performance in the form of RPC latency measurement to detect and troubleshoot networking issues that may affect Elastic Storage operation.
AFM parallel data transfers: Active file management (AFM) environments can now support parallel data transfers between sites. Now, multiple gateway nodes can be used to transfer a single file, greatly improving performance. Additional AFM enhancements: Elastic Storage 4.1 includes a number of features that optimize AFM operation. These include improved pre-fetch performance to pre-populate the AFM cache so file data is resident locally before it is directly requested by an application. To improve data transfer performance in the data center, AFM now supports the native GPFS protocol for AFM file transfers for improved integration of Elastic Storage features and attributes.
NFS data migration: Data migration eases data transfer when upgrading hardware or buying a new system. Elastic Storage data migration can minimize downtime for applications, move file data along with associated file permissions, and consolidate data from multiple legacy systems into a single more powerful system.
FPO enhancements: Elastic Storage File Placement Optimizer (FPO) enhancements include data locality-aware file system recovery and improved performance for concurrent directory changes to better support Hadoop and big data workloads.
A New Flash Paradigm: One of the coolest features of Elastic Storage 4.1 is the introduction of Elastic Storage Flash Local Read Only Cache (LROC). Flash LROCs are consumer grade Solid State Disks (SSDs) that are placed directly in the Elastic Storage clients, as shown in the diagram below.
And while there are many Flash-based products available to accelerate storage performance, Elastic Storage Flash LROCS are unique in that they place the flash cache across the “great network latency divide” from the shared storage device directly into the application node. This effectively removes network latency altogether for data access from a shared storage system – a pretty amazing feat.
This is in contrast to other shared storage systems from NetApp, EMC, Oracle and others that have similar caches, but they are placed across the network from the servers, in the shared storage system ‒ adding network latency. Elastic Storage Flash LROC is like having the performance boost of conventional server side flash caches with the manageability and data sharing benefits of a centralized shared storage system. What’s more, because the caches are read only, their access pattern maps to the wear characteristics of inexpensive consumer grade SSDs, so it’s an extremely elegant, cost effective way of boosting IOP performance for many applications.
Elastic Storage Flash LROCS also reduce network traffic substantially, benefiting performance across the board. And how fast is Elastic Storage Flash LROC? Benchmark tests have shown up to 6x IOP performance improvement over high speed disk access. This is shown in the figure below where, in this example, two consumer grade 200 GB SSDs cache a forty-eight 300 GB 10K SAS disk Elastic Storage storage system.
Initially, with all data coming from the disk storage system, the client reads data from the SAS disks at ~ 5,000 IOPS, but as more data is cached in the Flash LROCS, client performance increases to 32,000 IOPS while reducing the load on the disk subsystem by more than 95%. Elastic Storage Flash Local Read Only Cache is an exciting new paradigm in flash storage, offering the performance and low cost of server-side SSD flash caches, with all the manageability benefits of shared storage- a very unique and consequential combination!
Elastic Storage as Software Defined Storage: Software Defined Storage is one of the hottest topics in the industry today. No surprise, since storage that uses standard hardware with all the important storage and management functions performed in intelligent software provides significant technical and economic advantages for customers. And that’s exactly what Elastic Storage does.
As a leader in this space, you’ll be hearing a lot more about IBM software defined storage solutions moving forward. To get started, check out the recently published eBook Software Defined Storage for Dummies, which examines data storage & management challenges, and also explains the IBM approach to high-performance, cost-effective storage. For more information on Elastic Storage 4.1, visit the Elastic Storage IBM website.