Q&A with Imin Lee and Trent Steele

IL: AccelOps started in 2007 when my co-founder and I left Cisco to pursue a more holistic approach towards data center management.  We had both been at Cisco, initially as part of the network management business unit and then subsequently through our first company which introduced the MARS security event management appliance (acquired by Cisco in 2005).  AccelOps offers an integrated data center and cloud service monitoring solution that provides end-to-end visibility across performance, availability, security and change management from an infrastructure and service perspective.  It is available as a virtual appliance or SaaS for enterprise customers and service providers.

IL: In the last few years, the proliferated use of VMs in the datacenter for consolidation, as well as the ‘anywhere, anytime’ requirements for datacenter and IT Business services availability have added many dimensions and more magnitudes of complexity in the data center. In addition, today’s data center environment is extremely dynamic.  Applications are dynamically load balanced, connect to giga speed content switching and reference enterprise data stores to support high speed computing, as well as context switching.  Not only is the data constantly transmitted from one machine to another, but the image of the machine are constantly switching by vMotion for resource management. The guest OS (VMs) heavily depends on the ESX layer, while the ESX layer has virtual switching built-in and at the same time connects to outside content switching. In addition, the storage layer is playing a much more active role in computation, as now the VM images are stored in it and dynamically loaded into servers. In one single word: the line between application, OS, network and storage are becoming very blurred and it is significantly difficult to have just one single technologic silo view to understand the entire data center operation.

In addition, datacenter service reliability and availability assurance requirement introduces availability, performance and deviation time limits to identify, analyze and resolve issues - called the SLA.  Now extend the computing architecture, resources and applications into the cloud via private, hybrid or SaaS offering.  The service-level dynamics and controls become even more complex and add to operational and business risks.

So you can see that the environment is very different now than 10 or 20 years ago. In order to manage such a complex dynamic environment and meet SLA requirements, there needs to be new and demonstrably better way to holistically monitor and manage the data center environment.