A Multi-Dimensional, Multi-Layer approach with Performance As A Service™ using the Perfonics™ tool set

By | February 2, 2015

A Multi-Dimensional, Multi-Layer approach with Performance As A Service™ using the Perfonics™ tool set

A framework to perform holistic Performance and Capacity Planning in a cost efficient and hands-off manner…

While performing Performance and Capacity Planning related projects, we are most likely to focus on a single layer, especially with large enterprises where the organizational roles limit responsibility within domains of each layer. This always creates a disconnected picture especially during performance troubleshooting activities. Additionally, going across the various layers for Storage, SAN Switches, and Server OS requires multitude of software applications with the expectation that they are already licensed and installed and can be leveraged immediately on as needed basis during a performance troubleshooting exercise. Moreover, the key is to be able to create the linkages across layers for storage devices to troubleshoot performance metrics such as latencies and throughputs (IOPS, MBPS, service time etc.)

One of the approach to get around these issues is to leverage the performance and capacity data which may already be collected (e.g. sar, iostat, perfmon for servers or storage data from EMC, Hitachi, IBM etc.). The data can also be collected on an ad-hoc manner for performance troubleshooting using canned scripts on as needed basis. The advantage with this approach is that no pre-existing large scale software applications are needed to be pre-installed and there are no requirements for intrusive hooks such as agents and probes.

One additional issue that can be show stopper is even if there are expensive pre-installed applications collecting performance, configuration and capacity data, we still need a bunch of skilled resources to interpret the data and to create the relationships of various infrastructure components such as file-systems to OS level devices to SAN switch ports to Storage Array devices and Physical Disks. We will also need to map the business applications to these infrastructure resources. Obviously the pre-installed performance applications also needs the care and feed to manage the operations of the application on an on-going basis.

The approach taken by Interscape in their Storage Performance tool set called Perfonics™ fills these gaps in a unique manner. The Perfonics™ approach is to provide data collection tool sets, if needed, or to leverage the performance, configuration and capacity data already being collected at the customer site. This is done by ingesting the data into the Perfonics™ database, which also creates the linkages across the multi-layers of Storage, Switches and Server OS and is also go across the 3 dimensions of Configuration, Capacity and Performance in a single view. The Perfonics™ cloud is implemented in a fashion that it can be activated with a very simple setup instantly. Moreover, the Performance As A Service™ methodology includes a SME and a back-end Team that customizes the analytics for the client as part of the service.

The Perfonics™ toolset is leveraged to provide Performance As A Service™ on an ad-hoc or ongoing basis for small to very large multi-petabyte infrastructure. This approach has been used for projects of up to 30PB storage on over 200 storage arrays and over 5000 servers.

The Perfonics™ can be used as public cloud SaaS based application or be installed in the private cloud.

Please contact alok.jain@interscape.io for any questions or to find out more on this.