The IT industry is well aware of the performance advantages of flash for persistent storage and the increasing cost advantages of flash over spinning hard disks in terms of power, cooling, space, and consolidation of workloads which translates into lower administration and licensing costs. Earlier concerns about durability and maintenance costs, have been put to rest by usage data over the past 8 years. Many data centers are moving to all flash storage and Gartner predicts that 25% of all companies will be using flash for their primary storage by 2020.
Enterprises are adopting flash as one of the tools for digital transformation of their business. Flash helps to accelerate cost efficiency and time to market to transform IT operations and processes. It can increase customer loyalty by delivering a richer customer experience through responsiveness and real time analytics and unlock new revenue streams and new markets that were previously constrained by lack of performance, capacity, or cost.
While the addition of Flash can deliver the most benefits for applications that require high performance, going to an all flash storage infrastructure can deliver even greater benefits, through simpler management, lower maintenance, lower power, cooling and space, and lower overall costs, even for storage applications that may not benefit from higher performance. One of our customers in Europe was able to recover enough space through conversion to an all flash storage infrastructure that he was able to avoid the cost of moving to a larger data center with higher power input. A good place to hear about the latest developments in flash and customer experiences with all flash data centers is the Flash Memory Summit in Santa Clara, August 9-11.
One of the examples of an all flash data center presented at last year’s Summit was The COOP Group. Coop is a leading retail company in Switzerland with 2000 retail sales outlets in Switzerland and 199 wholesale and production sites in Switzerland and abroad. Their core business runs on SAP with Oracle on IBM p-Series servers with AIX. They are using SAP/HANA for real time business intelligence, and run other platforms on VMware and Solaris, Linux, etc. They have a DR requirement for long distance (120km) Asynchronous Remote Replication. Their current storage solutions were unable to keep up with increased demands. Nightly Business Warehouse consolidation of distribution center data could not keep up and the YoY growth of online retail applications was resulting in longer response times. They had insufficient visibility into storage health, making it difficult to diagnose performance issues and size latent demand. COOP adopted an all Flash strategy to solve their performance bottlenecks and reduce complexity. Their flash strategy required an always on infrastructure, with DR capability, and future proof technology that will easily integrate future flash technologies as they become available. The strategy also required management tools to provide the visibility to monitor and manage capacity and performance of storage /server workloads.
They chose a Hitachi VSP with 134 x 3.2 TB Hitachi Flash Modules with Universal replicator for asynchronous remote replication. For management of this storage environment, they are using the Hitachi Data Systems I/O portal, a completely cloud based tool to access capacity and performance graphs of their HDS systems without installing additional software on premise. With only a few mouse clicks they can generate the graphs revealing potential capacity constraints or performance bottlenecks of their Hitachi Data Systems storage systems. Dedicated graphs provide just the right level of information to quickly judge the workload, performance and impact of all connected servers.
Since this case study was presented, COOP has increased their Flash capacity to 1.65PB with the
purchase of our latest 6.4 TB flash modules with embedded compression. This further reduced their TCO for flash and validated the future proof capability of Hitachi Flash technology.
The results of choosing an Enterprise all flash infrastructure provides them with the following benefits:
Remote Asynchronous Replication – no distance limitation. Reduced costs by replacing host based replication. No host impact on response times, with RPO measured in seconds and RTO in minutes (depending on the application)
The new subsystem eliminated hours of elapsed time for critical applications. It reduced response time by a factor of 10 with average response times well below 1 ms for the entire subsystem.
Online customer satisfaction was improved with faster shopping cart checkout time.
Impact on their business warehouse not only insured that daily consolidation of distribution center data completes on time, but also enabled them to do a lot more analytic runs for better planning and decision making. With SAP/HANA they can do real time stock adjustment planning.
By migrating to an all flash subsystem they were able to reduce their infrastructure by 60%, from 5 to 2 racks, while increasing capacity by a third.
At this year’s Flash Summit, Dan King, IT director at Wellmark will be speaking about the financial and operational benefits that Wellmark realized with their VSP G1000 and Hitachi Accelerated Flash. His session, How to Fund Flash Storage will be presented in the Forum D-12 Enterprise Applications, Part 1, Tuesday August 9 8:30 t0 10:50 am.
Other sessions that you might want to attends are:
Title: All-Flash Data Centers—Emerging Use Cases
Presenter: Mark Adams, HDS Product Marketing Director
Forum D-12: Enterprise Storage Design, Part 1, Tues. Aug. 9th 3:20-5:45pm
Title: Accelerating Business Analytics with Flash Storage and FPGAs
Presenter: Satoru Watanabe, Research Engineer, Hitachi
Forum Q-21: Flash in Big Data and Analytics (Data Management Track), Wed. Aug. 10th 8:30-
Title: How Flash will transform enterprise applications
Panel Participant: Matt Pujol HDS Director Flash Storage Product Management
Panel 302-C: Thursday August 11 from 9:30 to 10:50
Please visit our booth and meet our engineers at booth number 811.