Virtualization Technology News and Information
Gear6 Unveils Enhanced Software Capabilities of CACHEfx Appliance to Better Analyze and Improve Data Center I/O Performance

Gear6, accelerating storage for real time application performance, today announced a powerful suite of tools to understand, improve, and manage I/O performance. Reflex OS, powering CACHEfx™ scalable caching appliances, provides a centralized view of data traffic between NFS clients and servers. As a result, system administrators now have unprecedented visibility into workload profiles to diagnose performance problems and also correct them. Typically, profiling tools are restricted to identifying performance issues but require separate problem resolution through manual intervention. With its combination of fine grained statistics and intelligent caching, the CACHEfx appliance uniquely integrates the ability to identify and fix I/O performance problems on the fly.

“The tightly integrated analytics of CACHEfx appliances provide deep insight into application I/O patterns from a central viewpoint without relying on agents,” said Martin Patterson, vice president of engineering, Gear6. “By combining a high level picture of application behavior plus a detailed NFS operations breakdown, administrators can quickly diagnose and resolve storage performance issues.”

Reflex OS provides a comprehensive view of application behavior and workload statistics that have previously been difficult or impossible to collect. Primary metrics include latency, I/O Operations per Second (IOPS), and bandwidth consumption. CACHEfx appliances record specific and granular application data requests, giving administrators an understanding of the active data set size, requested files, and cache hit rate. While other analysis tools often layer onto existing systems, adding overhead and potentially disrupting the environment, Reflex OS analysis features are built into the software, seamlessly providing insight while also dramatically increasing I/O performance.

“Gear6 has delivered real value for customers to observe and automatically manage both dynamic and persistent data with their CACHEfx scalable caching appliances,” said Steve Duplessie, founder and senior analyst for Enterprise Strategy Group. “The innovative centralized caching approach allows dynamic data to remain in high-performance cache, yet also retains existing persistent, disk-based storage to cost-effectively scale capacity for large content-centric data stores. This is having your cake and eating it too."

CACHEfx appliances deploy as a shared network resource, providing end-to-end visibility for all accelerated NFS servers and clients, aggregating I/O profiles across multiple applications, storage devices, and up to thousands of clients. This unique visibility is one of the reasons customers describe part of the Gear6 solution as ‘the best NFS sniffer on the market.’

Modern applications frequently support thousands of users and rely on large server farms accessing data from multiple storage systems. These large scale shared I/O architectures present a significant performance improvement challenge to system administrators. By caching most frequently accessed data in large scale memory, CACHEfx appliances instantly and seamlessly accelerate data access to existing storage devices. Files automatically populate the cache upon application request without manual data migration or complex multi-tier approaches. Real time statistics provide live performance and traffic monitoring enabling system tuning if needed. As applications evolve or are added over time, CACHEfx supports live configuration adjustments to optimize the changing environment.

Centralized storage caching solutions support any NFS-based file system, including clustered file systems. They attach transparently to the network using standard Gigabit or 10 Gigabit Ethernet connections and no changes are required for existing applications or systems. Gear6 provides a straightforward Management Information Base (MIB) and supports the Simple Network Management Protocol (SNMP) so customers can tap into the rich statistical data and integrate with existing enterprise management frameworks.

Published Monday, January 28, 2008 7:51 PM by David Marshall
There are no comments for this post.
To post a comment, you must be a registered user. Registration is free and easy! Sign up now!
<January 2008>