Hitachi Unified Storage and Hitachi NAS Platform Performance Optimization with Flash Acceleration

Editor's Notes

• #6: Several system and application drivers are increasing the demand for high-performance storage solutions going forward
• #7: VMware – VMFS file system (Server consolidation + VDI)VMFS on block storage requires large amounts of random IOPSVMFS and customers can have higher density of VM’s per datastore thus reducing disk wastage.Customers can leverage vSphere 5 features like SDRS for tiered storage datastores for movement of VM’s between (tiers) using vmotionCustomers can leverage all vStorage API’s that are currently supported on the VSP i.e. VAAI, VASA, SRM & MPPRed Hat (OLTP workloads) – GFS / OCFS file systemsOLTP workloads demand high read/write IO with low latencyFor an I/O constrained OLTP database, many companies add HDDs to separate the database across many drives to alleviate the I/O bottleneck.Adding more HDD for just IOPS and not for increase disk capacity increases cost.IBM - GPFS file system (Analytics + Data Mining etc)Customers use GPFS to do low latency analytics, high-speed back up and recovery, high-speed file ingest, and optimize data workflow, enabling faster business process performance. Flexibility to support a wide variety of platforms - Linux, AIX, Windows
• #8: And the reasons are clear. The spinning disk performance of yesterday has been leapfrogged by the flash storage performance of today. In fact, on a device-by-device basis, a single Hitachi Accelerated Flash storage module that we deploy in our storage systems is capable of about 500 times more operations per second than what we once called a ‘high performance’ disk drive.
• #9: This chart demonstrates the per-device expectations of common storage media, and shows how significant the technological leap is to a Hitachi Accelerated Flash storage module. Any SSD leaps over the performance bar set by spinning disk drives, but Hitachi Accelerated Flash soars further ahead providing 5X more IOPS than an industry-standard SSD at similarly astounding response times. All this is done while offering significantly more capacity per device and the advanced memory management expected in such a robust Hitachi design.
• #10: One of those key innovation areas is now available to users in what we call Hitachi Accelerated Flash (HAF) storage, and our own flash storage memory controller that we internally called “Ashigara” and the unique packaging that it offered us. With over 80 patents either granted or pending for flash around the designs for this purpose built flash storage, it enables us a leadership position in flash.Built from the ground up to enable cost-effective, high-performance and high-density deployments of flash storage in enterprise data center environments, our Accelerated Flash storage raised the bar high above the capabilities of industry-standard SSDs that are the norm in the industry.The ability for Hitachi Accelerated Flash to manage more flash per controller, provide high-bandwidth performance and run advanced wear-leveling and diagnostics means our customers get the flash capacity and performance they need, without sacrificing reliability.
• #11: But Hitachi flash innovation extends well past our Flash controller and up into our HUS VM and VSP systems. While the SPC performance numbers that Silverton Consulting mentioned earlier in the presentation already demonstrated the high performance of our HUS systems, the architectures of the VSP and HUS VM allow for even greater performance scalability and customization.For years, storage systems were built with the idea that the performance was held back by spinning disk drives, and system software was optimized to make the most out of that situation. With flash, vendors are rushing to re-architect key parts of their onboard system code to more quickly move data around the inside of storage systems to keep up with the high performance of Flash. Hitachi, however, isn’t scrambling to deliver this customization. We already have.VSP optimization was delivered in 2012 and HUS VM will be delivered in June 2013. The system code for these arrays see significant flash-focused re-engineering, with over 30 fundamental changes to the code to enable higher thread-count processing and faster movement of data. Response times improve significantly and provisioning and functions become twice as useful.Of course, in staying with Hitachi’s focus on high-availability, these changes are non-disruptively installed and become transparent to other installed storage replication or management tools.
• #12: A layer above the infrastructure, sits a family of offerings aimed to provide intelligent access to business content and information. These software-based products provide management, protection, archiving and searching of your files, objects and information across the complete lifecycle.
• #13: A layer above the infrastructure, sits a family of offerings aimed to provide intelligent access to business content and information. These software-based products provide management, protection, archiving and searching of your files, objects and information across the complete lifecycle.
• #16: Such high performance optimizations allow us to produce no-compromise all-flash storage performance in our HUS VM, which competes with any all flash array for high performance with sub-millisecond response time to 500,000+ IOPS, but still brings all the functionality, availability and flexibility benefits of a non-Flash HUS VM. All of this happens with no changes to the serviceability and manageability of the systems. Hitachi Command Suite manages the system and the Dynamic Provisioning, Thin Image and the full suite of Hitachi replication tools can be leveraged as required.What if you decide that you want to add tiering and cost-effective hard drive capacity? HUS VM can be upgraded to a tiered storage system.Want a unified SAN/NAS system? No problem.
• #17: Our king of the hill is our multi-controller, massively scalable and rock-solid VSP. It has already proven 1M+ IOPS and the ability to handle multi-platform loads including the mainframe. The optimized code for VSP helped grow the performance scalability multiple times ahead of where the product was first launched delivering significant value to our customers.Many VSP customers are deploying Flash and tiering to scale the performance of their applications without having to completely re-architect their environments. Advanced management tools enable tiering but still provide a level of control to manage quality of service across applications by defining how much cache applications can utilize and even to “pin” critical data in flash.The other leading argument for tiering is an economic one.
• #23: Real customers are getting theefficiency of tiered flash storage today. Turkcell, a large GSM operator in Europe, is one such example and this is really a great case study that you could spend some timewith, seeing how the power of storage virtualization really can change an IT transformation. Turkcell leveraged flash and tiering to keep mainframe performance high, while vastly shrinking the amount of disk drives and racks deployed in their infrastructure.
• #24: Several system and application drivers are increasing the demand for high-performance storage solutions going forward