HPC Job Bank
HPC in the Cloud


Dedicated to covering high-end cloud computing
in science, industry and the datacenter

Language Flags

Making Storage Access Application-Centric


In the world of high-performance enterprise computing, it is generally understood that mission-critical applications need on-demand access to the database. The higher the latency, the less effective the solution. Not as much is said, however, about the storage layer residing below the database. Isn't it important that applications have as-needed access to the underlying storage, as well?

Pillar Data Systems is trying to remedy this situation with its Axiom family of hardware and software products, which allow users to dynamically provision storage to meet the service levels required by specific applications -- something Pillar calls "application-aware storage." Essentially, according to Pillar's director of product management, Rob Commins, the company is creating a way for the storage platform to get allocated to applications the same way virtualization software allocates server space or grid middleware allocates computing resources. "All applications are not created equal," said Commins. "Therefore, not all storage should be created equal."

Using a method called "statistical short stroking," as well as thin provisioning, Axiom allows more-important applications to access the outer portion of the disk (which results in faster access) more of the time, while still allowing the application to access lower-priority, lower-speed stored data on the inner portions of the disk the rest of the time. These levels can be changed as needed when new applications enter into the Axiom storage environment. Traditional storage platforms, explained Commins, do not allow users to allocate a disproportionate amount of resources to particular applications, thus negating any "elasticity" that customers might be trying to drive. Additionally, added Commins, Axiom's methods lead to maximum utilization rates: Axiom users average between 60 and 80-plus percent utilization, whereas traditional systems average between 30 and 40 percent. These improvements in performance and utilization are necessary, says Commins, because while the storage industry has done a good job of figuring out how to handle data growth and scalability, "We're seeing application proliferation as one of the biggest headaches on the storage side."

Another big headache, however, is less obvious -- and even less foreseen by companies thinking they're solving all of their problems by embracing server virtualization. Virtualization solutions like VMware, says Commins, bring back a whole host of problems in terms of performance, which ends up having a negative effect on ROI. Virtualized environments cause spindle contention is disk drives, which, when addressed using traditional techniques, requires either a huge sacrifice in performance or a significantly larger amount of storage -- typically 20 to 200 percent more than needed. "So," Commins explained, "the guys holding the wallets ... are sitting there surprised, going, 'Wait a second, I thought this virtualization thing is supposed to save me economic and operational costs, but all that's really happened is we've shifted server dollars back into storage dollars.'" Pillar has addressed this problem by creating profiles for use in virtual environments -- including VMware and Oracle VM -- thus bringing Axiom's ability to carve out separate volumes with separate storage services to each VM.

Cost-wise, said Commins, Axiom really starts delivering savings when customers combine applications on the platform. Because most storage platforms essentially utilize the same components, Axiom probably costs about the same as others when used with only one application, "but as customers leverage the architecture for what it's built for, we see anywhere from 20 to 80 percent advantages in dollars per gigabyte and dollars per IOPS versus traditional methods.” Axiom also saves customers money when compared to traditional separate database and storage layers -- which generally consist of two systems from two different vendors, and where resources can't be allocated between the two -- because you can allocate performance and capacity when necessary, said Commins.

Axiom's application-aware storage is a "necessary" and "very smart" technology, believes Deni Connor, principal analyst with Storage Strategies Now, because users are constantly tying storage to applications, database administrators are always asking for more capacity, and users are always complaining about performance. Outside of the mainframe space, she said, "For a company to actually go in and affect how those storage resources are used, how performance is built into it, how cache memory and disk are built into it, is something that hasn't really been done before."

As a result, said Connor, customers with applications and databases that need fast access to the storage layer, as well as those running HPC environments, will likely be clamoring for solutions like Axiom. "Everyone talks in terms of the application and what they need and when they need it and what type of service level is attached to it," she noted. In the two years Pillar has been marketing and selling Axiom, the company has accumulated more than 350 customers running a total of more than 550 Axiom systems, with most of the uptake coming from the financial services, medical, education, government and, as of late, legal markets.

Whereas other vendors have tried to do some of what Axiom does by adding optimizers or additional layers to their offerings, Commins believes Pillar has attracted so many customers because its approach of working the concept of application-aware storage into the every aspect of the Axiom architecture makes it so effective and so unique, not to mention cost-effective. You can derive some significant ROI, he said, "from investing in a platform versus yet another mutually exclusive storage system in the bone pile of gear on the datacenter floor."

Most Read Blogs

HQuanta

Short Takes

Running Computational Fluid Dynamics in the Cloud

May 16, 2013 | When it comes to cloud, long distances mean unacceptably high latencies. Researchers from the University of Bonn in Germany examined those latency issues of doing CFD modeling in the cloud by utilizing a common CFD and its utilization in HPC instance types including both CPU and GPU cores of Amazon EC2.
Read more...

In Support of Cloud-based Rendering

May 10, 2013 | Australian visual effects company, Animal Logic, is considering a move to the public cloud.
Read more...

Internet2 Awards Program Seeks Innovative Applications

May 10, 2013 | Program provides cash awards up to $10,000 for the best open-source end-user applications deployed on 100G network.
Read more...

HPC and the True Cost of Cloud

May 08, 2013 | For engineers looking to leverage high-performance computing, the accessibility of a cloud-based approach is a powerful draw, but there are costs that may not be readily apparent.
Read more...

Sponsored Whitepapers

Best Practices in Big Data Storage

05/10/2013 | Cleversafe, Cray, DDN, NetApp, & Panasas | From Wall Street to Hollywood, drug discovery to homeland security, companies and organizations of all sizes and stripes are coming face to face with the challenges – and opportunities – afforded by Big Data. Before anyone can utilize these extraordinary data repositories, however, they must first harness and manage their data stores, and do so utilizing technologies that underscore affordability, security, and scalability.

Exploring the Potential of Heterogeneous Computing

04/02/2012 | AMD | Developers today are just beginning to explore the potential of heterogeneous computing, but the potential for this new paradigm is huge. This brief article reviews how the technology might impact a range of application development areas, including client experiences and cloud-based data management. As platforms like OpenCL continue to evolve, the benefits of heterogeneous computing will become even more accessible. Use this quick article to jump-start your own thinking on heterogeneous computing.

Sponsored Multimedias

Newsletters

Stay informed! Subscribe to HPC in the Cloud email Newsletters.

HPC in the Cloud Update
HPCwire Weekly Update
Digital Manufacturing Report
Datanami
HPCwire Conferences & Events
Job Bank
HPCwire Product Showcases



HPC Job Bank


Featured Events



  • June 16, 2013 - June 20, 2013
    ISC'13
    Leipzig,
    Germany

  • June 17, 2013 - June 18, 2013
    Forecast 2013
    San Francisco, CA
    United States




HPC in the Cloud Conferences & Events