Software Defined Storage Meets Software Defined Networking

Coho Data Delivers Scalable Multi-Purpose Storage Platform

One of the challenges facing data centers today is the diversity of the applications they have to support. They range from the virtualized infrastructure, with its highly random I/O pattern, to the transactional I/O of databases, to the high capacity reality of long term data storage. Many data centers trying to tackle these disparate I/O challenges, despite the expense of acquiring and managing them, often resort to separate silos of storage. Coho Data has recently launched a new storage solution that leverages both software-defined storage (SDS) and software defined networking (SDN) to deliver a system capable of handling a wide variety of workloads.

Hardware Granularity

The key to Coho’s ability to support mixed workloads comes from its ability to granulize most of the storage functions. The first building block of the solution is the MicroArray, which is perfectly configured to deliver maximum performance for its configuration. Each MicroArray has a combination of PCIe Flash, 10GbE Networking, Hard Disk Storage and CPU processing power. There are two 800GB PCIe Flash SSD per MicroArray. Coho is using the Intel 910 PCIe SSD; which is one of the highest performing and most affordable PCIe SSDs we’ve seen. For capacity, each one of the MicroArrays has 6x3TB hard disk drives. Finally, each MicroArray has dual 10GbE/1GbE ports for connectivity. MicroArrays are sold in pairs per chassis for redundancy.

Software Granularity

On each of the MicroArrays, Coho runs a software defined storage stack that manages the resources of each array, while also providing interaction with the other MicroArrays in the configuration. Each MicroArray can make decisions on data placement and data protection requirements. Because of the hardware granularization, it has the processing horsepower to run sophisticated, use case based analytics to ensure that the right data is on the right storage type as often as possible.

Data protection is done through replication instead of disk RAID using a RAIN approach. Data starts off in flash and then is copied to other MicroArrays based on the given data protection policy created. The secondary copy, however, can reside on hard disk while the active part of the primary copy is retained on PCIe Flash. This provides rapid response time to mission critical applications while providing cost effective availability. Rebuild times for failed hardware is much faster than with traditional RAID as well because all the MicroArrays in the system participate in the rebuild.

Finally, due to its data hypervisor capabilities, the storage software has the processing power to fully consume the 10GbE ports available to it. This allows Coho to deliver as much as 180k IOPS per MicroArray (random 80/20 r/w, 4k block size workload) and the linear scalability to provide over a million IOPS.

Protocol Granularity

The Coho DataStream system also provides protocol granularity. Today includes VMware focused NFS but next year, the list will quickly expand to include general purpose NFS and SMB. In the future, specific protocols for Mongo, Riak, Hadoop and even direct storage system access will be available. This protocol granularity further supports Coho’s claim of a single solution for all workloads.

Software Defined Networking

One of the key challenges facing scale-out systems, especially ones that leverage PCIe SSD for performance, is the latency caused by inter-nodal communication. Instead of interconnecting the MicroArrays over traditional storage interconnects, like SAS or iSCSI, Coho leveraged SDN to significantly reduce this latency.

MicroArrays are aggregated using the OpenFlow-enabled DataStream10GbE switch, with support for an active/active redundant configuration. The SDN features on the DataStream Switch enable key storage system logic including data placement, routing and load balancing to be delivered from the network itself. This innovative use of SDN enables the system to maintain low latency as it scales from one to dozens of MicroArrays.

Storage Swiss Take

Two opposing forces – the need for high performance, low latency storage and the need for modest performance, high capacity, highly durable storage are polarizing the data center. Polarization is actually popular with vendors because it seems to justify the need for multiple system purchases. Coho may set that vendor strategy on its ear and provide IT Planners with what they want; a single platform that can truly manage multiple workload types. If successful, Coho could dramatically reduce the cost to acquire and manage high performance and high capacity storage.

Coho Data is not a client of Storage Switzerland

Unknown's avatar

George Crump is the Chief Marketing Officer at VergeIO, the leader in Ultraconverged Infrastructure. Prior to VergeIO he was Chief Product Strategist at StorONE. Before assuming roles with innovative technology vendors, George spent almost 14 years as the founder and lead analyst at Storage Switzerland. In his spare time, he continues to write blogs on Storage Switzerland to educate IT professionals on all aspects of data center storage. He is the primary contributor to Storage Switzerland and is a heavily sought-after public speaker. With over 30 years of experience designing storage solutions for data centers across the US, he has seen the birth of such technologies as RAID, NAS, SAN, Virtualization, Cloud, and Enterprise Flash. Before founding Storage Switzerland, he was CTO at one of the nation's largest storage integrators, where he was in charge of technology testing, integration, and product selection.

Tagged with: , , , , , ,
Posted in Briefing Note

Enter your email address to follow this blog and receive notifications of new posts by email.

Join 17.4K other subscribers
Blog Stats
  • 1,979,429 views