USE CASES

Breaking free from the limitations of rotational storage, WhipTail’s flash-based storage array offers your “hot-spot” application sets incredible speed, throughput and cost savings. Both random and sequential operations are subject to constraints presented by traditional storage solution latency. However, WhipTail’s flash-based storage array, the XLR8r, can deliver 250,000 random write IOs with sub-millisecond response times, ensuring that neither IO contention nor latency become a costly issue.

The XLR8r’s Racerunner operating system is optimized to manage all writes sequentially, generating unparalleled sequential write performance and increasing the endurance of the SSD drives populating the XLR8r’s arrays. Enterprises utilizing our flash-based array realized tremendous productivity increases in their complement of applications that are the hardest to improve:

 

ORACLE & SQL

Oracle and SQL are typically adopted by organizations that require a high-performance and / or large-scale database solution. Oracle & SQL DBAs and storage administrators are likely familiar with the significance of storage design in relation to the performance and success of an Oracle implementation, particularly a clustered deployed intended to provide high availability and business continuity. In many cases, the sluggish performance of an Oracle DB heavy transaction windows or the execution of reports with lengthy response times stems from IO-related bottlenecks. Organizations aiming to increase productivity and quell dissatisfied end-users by accelerating query and report completion times must consider the IO/s performance and latency of the storage solution hosting its data.

To understand the interplay of IOPS and latency in relation to database performance, it’s important to comprehend how Oracle manages IO. Oracle stores all logical databases structures, including tables and indexes, in physical datafiles. Oracle datafiles are usually accessed in random order; however, Oracle optimizations provide for caching of data in datafiles into memory as needed, which will increase the random IO performance of datafile transaction within limits. Redo logs (or transaction logs for SQL), which contain a record of all transactions for recovery purposes, and archived redo logs are written in sequentially. While sequential writes are less demanding on traditional disk arrays, an excessive number of mechanical spindles are still necessary to sustain sufficient IOPS to meet the needs of high-IO implementations.
Back to the top.

 

Server Virtualization

With the proliferation of server virtualization beyond test and development environments over the past five years, a new and more challenging obstacle is emerging—storage sprawl. Server Virtualization involves users competing for the same resources, which is especially problematic for mid-to-large organizations’ virtualized environments. This is when disk performance, CPU and memory issues arise.

With multiple systems contending for the same disk supply, sequential workload can turn into very demanding random workloads, which is extremely difficult for hard disk drives to handle efficiently. When functioning on an HDD storage array, virtualized servers cannot transfer files without being interrupted by another file transfer. As a result, enterprises deploy entirely new storage infrastructures to meet the IO demands to continue to scale-out their virtual environments.

As a completely flash-based storage array, the XLR8r enhances virtualized environments, sees sub-millisecond data access time and random and sequential workloads are no different to the XLR8r. Virtualized applications can finally achieve the performance required by enterprise IT departments and end-users.
Back to the top.

 

Virtual Desktop Infrastructure (VDI)

Considered the “nirvana” of IT computing, the single biggest challenge facing enterprises that wish to deliver a VDI infrastructure is the overwhelming storage infrastructure required to support it. It’s not physically or financially practical to attempt this kind of paradigm shift in computing on legacy rotational media. By deleting 90% of these hard and soft costs, the XLR8r removes this obstacle by focusing on what a VDI environment craves: high amounts of WRITE IO. With 250,000 IOs packed into a small 2U array that is fully populated by cost-effective MLC flash drives, the XLR8r is the missing link in your virtualized architecture.
Back to the top.

 

Software Development Life Cycle (SDLC)

Development and Test Environment Deployment is a major bottleneck in Software Development Life Cycles (SDLC). WhipTail’s XLR8r, an entirely flash-based array, improves productivity in SDLC by dramatically reducing time to generate infrastructure, operating systems, and test databases.

The creation of large integrated SDLC environments depends on the rapid creation of hundreds and sometimes thousands of VMs, databases, and runtime environments. Building these test databases has become the primary bottleneck, and traditional data storage cannot handle the rapid provisioning and creation of large, secure test databases.

The XLR8r has reduced test database creation times by 90% with its extreme Data Write performance of up to 1.9 gigabytes per second and the ability to process and absorb 250,000 WRITE IO per second. Even the largest test database extracts and copies can then be created in hours instead of days, dramatically reducing cycle times to test new software functionality.
Back to the top.

 

Code Coverage and Analysis

Like any I/O demanding application, Code Coverage and Analysis cannot meaningfully improve performance with traditional HHD storage, regardless of how many terabytes of it are purchased. Because of these limitations, the process of moving large amounts of data from traditional data storage devices to host servers running the Code Analysis is unbearably slow, and negatively affects productivity. Neither an increase in the amount of DRAM Cache nor a faster network connection will correct this.

Some vendors may recommend “Short Stroking”—buying substantially more capacity than is needed and placing data on the outside tracks of disks. While Short Stroking does moderately improve performance, a 200 terabyte system short stroked for performance can yield only 40 usable terabytes, and it creates only a minute increase in actual storage subsystem performance. A stop-gap scenario that is now old in the tooth.

WhipTail offers industry standard, NAND flash-based storage arrays, which can support numerous RAID levels, standard Fibre Channel, iSCSI, IB-QDR, NFS, and CIFS/SMB. The XLR8r works with all existing infrastructure, including all standard multipathing software and standard server operating systems.
Back to the top.

 

Email

The most critical business application ever created, bar none is email. It’s the lifeblood. Yet, despite enhancements to Exchange 2007 and 2010, storage performance continues to be essential to the success of a deployment. Transactional IO generated by end-user activity is critical to the overall design as the IO latency can have a major impact on the user-experience. Operations generated via online Exchange clients (sending, receiving, deleting, sorting) in Outlook can create IO latency that impacts all users.

The WhipTail XLR8r, capable of 250,000 IO/s simplifies storage planning for Exchange by eliminating the need for involved calculations to determine IO/s and spindle requirements. Thus, providing organizations with the high-speed and cost-effective performance required. Nobody ever complained, “email is just too fast today.”
Back to the top.