Interview: Diablo Technologies Disrupts with Memory Channel Storage

Print Friendly, PDF & Email

Given the massive amounts of data generated daily, IT Managers have found a need to link data directly to their managing applications. Diablo Technologies has done just that via the revolutionary technology known as Memory Channel Storage. We caught up with Riccardo Badalone, CEO and Co-Founders of Diablo, to shed more light on this disruptive approach.

insideBIGDATA: I understand that Diablo Technologies does Memory Channel Storage (MCS). What is this exactly?

Riccardo_BadaloneRiccardo Badalone: Memory Channel Storage is an innovative storage architecture that enables non-volatile media to reside within the memory subsystem. Today, MCS modules are enabling NAND flash to directly interface with the CPU via DDR3 at near-DRAM speed. This revolutionary approach dramatically minimizes storage latency and provides a level of latency determinism that is unachievable for alternative flash storage solutions (i.e. via PCIe, SAS, or SATA). In addition, unlike those alternative solutions, MCS does not require a trade-off between latency and I/O throughput. By leveraging the native parallelism of the memory subsystem, MCS provides scalable performance while simultaneously providing ultra-low response times. Via strategic partnerships with SanDisk and IBM, the first MCS-based products have now been publicly announced (SanDisk™ ULLtraDIMM™ and IBM eXFlash™ DIMM).

insideBIGDATA: Who out there should be paying attention to this approach?

Riccardo Badalone: Anyone using or designing Enterprise storage solutions should be paying attention. The unique performance and scalability characteristics of Memory Channel Storage are truly disruptive and will both accelerate existing applications with no changes required and enable new use cases.

insideBIGDATA: You’ve recently announced VMware ready status. How did this partnership come about and what does it mean in a technological sense?

Riccardo Badalone: We are very proud of our strategic partnership with VMware and our recognition as VMware Ready.  The partnership came about as the result a long collaboration at both the executive and engineering levels.  Technologically, there is a natural synergy between the capabilities offered by Memory Channel Storage and the requirements imposed by virtualized workloads. The internal test and benchmark results we’ve seen thus far are very compelling and we’re extremely excited about the benefits we can offer in for both server virtualization and VDI applications.

insideBIGDATA: Let’s talk a little bit about MCS and Big Data. So, Big Data–hype or here to stay?

Riccardo Badalone: Big Data as a class of datasets is definitely here to stay. The ever-increasing rate of data creation and the growing desire and requirements for collection, categorization, and analysis will make sure of that.

insideBIGDATA: What can MCS do for data analytics in the enterprise realm?

Riccardo Badalone: Because of its disruptive performance characteristics, MCS is uniquely suited to expand analytics capabilities.  By eliminating the bottlenecks currently imposed by traditional storage solutions (even flash-based ones), MCS enables applications to access and, hence, analyze data more rapidly. In addition, MCS will enable system architectures that facilitate more effective analytics.  For example, in-memory analytics are critical in the enterprise because some applications simply cannot tolerate the huge discrepancy between RAM access time and storage access time. With Memory Channel Storage, that performance gap is dramatically decreased….and with consistent performance. No longer constrained purely to RAM capacity, analytics engines can now have access to terabytes of deterministic, ultra-low latency storage….opening up possibilities that previously did not exist. Real time analysis is also a tremendous use case, where very large sets of data must be analyzed and stored (or written) very quickly. With such low, predictable write latencies, MCS is a perfect solution this type of use case.

insideBIGDATA: What does Diablo have in store for the future especially relative to Big Data.

Riccardo Badalone: The future of Diablo and Memory Channel Storage is very exciting. The interest in and uptake of our initial products has been amazing and we are committed to fully realizing their potential in real-world deployments…this definitely includes applications that collect, manage, and analyze Big Data. That said, we are not slowing down for a moment. The first generation of Memory Channel Storage is only the beginning. There are future advancements in both hardware and software that will continue accelerating Big Data applications and improve system efficiency even further. Stay tuned.

 

Sign up for the free insideBIGDATA newsletter.

Speak Your Mind

*