Based on QLogic’s Mt. Rainier technology, the FabricCache 10000 Series adapter combines a Fibre Channel host bus adapter (HBA), caching and I/O management with connectivity to a server-based PCIe flash card. Expected to ship at the end of the month, the company said it is exceptionally simple to deploy and manage and transforms single-server, captive cache into a shared, performance-enhancing resource for a wide range of enterprise applications that rely on SAN storage. Mt. Rainier, which debuted last September, is a Server based Storage Accelerator (SSA) which captures all IO and redirects it to a flash-based cache,
It seems like everybody wants to grab a share of the server-based flash market that has been dominated primarily by Fusion-io (ioScale). In just the last week we’ve had announcements from Sanbolic (Melio5), Violin (Velocity), and storage’s dominant vendor EMC (XtremSF). Commenting on ‘Flash Announcement Week’, Mark Peters, Senior Analyst, Enterprise Storage Group, said that you need both management and table-stakes storage functionality if flash is to be more than just a turbo-booster used for limited and specific workloads.
QLogic said the market has been looking for the next level of server-based caching. “A very large number of enterprise applications are in these clustered environments,” said Cameron Brett, Director of Solution Marketing. “If cache is limited to a single server, you have a bottleneck there.”
QLogic has been innovating at the fabric layer for more than two decades, so it was no surprise that they looked at a new type of HBA as the way to solve the problems with caching 1.0, said analysts at Storage Strategies NOW. Challenges with caching 1.0 – keeping hot data on the flash card and going to shared storage for everything else – include the unavailability of cache when a virtual workload is moved to another physical server, support for clustered applications, and the requirements for configurations and scripts. ‘QLogic’s FabricCache solves these issues and creates server-based caching 2.0.’
The FabricCache QLE10000 solution uses two PCIe slots. The first slot houses the adapter which includes the System-on-Chip (SOC) – which is responsible for finding a given block for read operations and will get it from flash if it exists on the local flash card or any other FabricCache card in the fabric – and is connected to a second PCIe board with a ribbon cable that operates as a PCIe interface and contains either 200GB or 400GB of single level cell (SLC) flash memory. The second slot provides power to the flash board.
‘This is server-based caching 2.0. Because the FC fabric has the bandwidth, and the SOC has the brainpower, reads from the storage area network (SAN) only occur if the block does not exist in any of the attached flash memory. This solves the problem of moving virtualized workloads between physical machines, because the workload has access to its original cache no matter to where the VM is moved.’
Storage Strategies NOW believes QLogic’s unique approach is likely to cause many OEMs and data center managers to rethink the deployment model. The ease of installation and support of the FabricCache solution is a major advantage over other methods, and the price point competes well with the combined pricing of HBA, flash card and cache software from other vendors.
Jeff Byrne, an analyst with the Taneja Group, is excited about the new offering. ‘FabricCache is the industry’s first adapter offering that enables the cache from individual servers to be pooled and shared across multiple physical servers.’
According to a new survey of nearly 280 IT managers and executives, more than 75% said that they are thinking about deploying a storage acceleration solution of some kind, and 56% indicated that server-based acceleration is one of the acceleration approaches they most value, with users overall valuing it more highly than array or network based approaches. But in contrast, only 26% said they had deployed a server-based acceleration solution to date, citing a number of obstacles that have prevented them from moving forward.
Taneja believes that FabricCache adapters will help customers to overcome many of the limitations of today’s server-based caching solutions, including eliminating the need for costly overprovisioning by allowing single-server flash cache to be transparently and coherently shared across multiple physical servers, irrespective of the operating systems, hypervisors and applications they’re running. ‘In effect, this enables FabricCache to serve as a caching SAN adapter, the first of its kind in the industry.’
FabricCache also provides economic benefits, noted Byrne. Because it can be integrated non-disruptively into any storage and server infrastructure, running any combination of hypervisors, operating systems and applications, FabricCache eliminates the costs of additional drivers and caching software, and helps to extend the life of an existing SAN.