Imin Lee, CEO of AccelOps (www.accelops.net), says:
Computing as a utility remains the dream of data center managers, where more resources are just available when you need them, like electricity when you turn on the lights.
Dynamic hosting and cloud computing are the core technologies that are going to make that dream of utility computing a reality.
In dynamic hosting and cloud computing environments, new network, server, VM and applications can be provisioned dynamically, by customer and at various time periods. That certainly sounds like utility computing, but as any IT manager who’s done it can tell you, it is far from throwing a light switch.
Certainly the first step is to position the virtualization and network elements needed to power the delivery of on-demand provisioned IT services, and then to integrate them with existing services, topics that get a lot of attention in the media.
What is not talked about, however, is that successful cloud adoption will depend on assessing data center and cloud management controls that will ensure a more reliable and secure delivery of private/ hybrid cloud services.
Managing security and compliance risks will be paramount, and equally critical is performance
monitoring.
Since service providers offer SLAs to meet availability and performance requirements, they have
implemented data center and cloud monitoring tools. Unfortunately, the same capability that is
bringing us closer to utility computing — dynamic provisioning — can impact the capacity of monitoring platforms, which can fail or drop operational data from being processed. This presents operational risks and potential contractual liabilities.
In dynamic cloud environments, you or your providers may not be able to adequately anticipate the operation data burst rates and load changes. So you potentially must bear additional license and capital expense costs to support excess monitoring capacity. Or you need to re-assess detailed monitoring capacity and added costs as you add new clients or resources, even if for a short period of time, which impacts the speed to deliver contracted services. This monitoring capacity limitation also adds administrative burden to build out and maintain the different monitoring components to monitor the dynamic resources.
That doesn’t sound like turning on a light switch to me. In order to achieve the dream of utility computing you need to be elastic, and to pay only for true usage. And your management software and hardware needs to be able to scale up and down as rapidly and dynamically as your computing resources to. Management and monitoring capability needs to change according to the data center and cloud environment’s computation power: network traffic, flows, packets, logs/events, metrics, and devices. Therefore, being able to dynamically scale is a prerequisite of being able to elastic.
A recent visit to a global Fortune 100 enterprise highlighted the difficulty of monitoring hybrid private/cloud environments with conventional solutions. Simply having multiple copies of a SIEM appliance,
whether virtual or otherwise, is not synonymous with being able to scale out event cross-correlation capabilities. True elasticity requires clustered cross correlation capabilities, dynamic deployment of virtual appliances for more processing power, usage accounting, and a flexible licensing model.
This capability is what I call elastic monitoring-the ability to dynamically scale monitoring resources and keep pace with dynamic virtual and cloud environments. Our company AccelOps, for example, provides this unique capability in our Integrated Data center and Cloud Monitoring solution. It seems we’re the first to market with this emerging technology, according to a Frost and Sullivan research report published in January 2011, although others will certainly attempt to follow us rapidly.
To conclude, data center managers planning a hybrid private/public cloud implementation need to
pay equal attention to security and monitoring requirements, and should look for a solution that can dynamically scale alongside their computing environment. It will also bring benefits in security, as new integrated systems include SIEM capabilities and further the convergence of SOC and NOC, as well as significantly greater automation in managing compliance reporting and investigating incidents.
Achieving this will take you one giant step closer to fulfilling the dream of utility computing.
Imin Lee is the founder and CEO of AccelOps, the integrated data center and cloud monitoring leader. Previously she founded Protego Networks, a leading provider of Security Information Management appliances acquired by CISCO (MARS), and served as the Director of Engineering in the Cisco Security Technology Group.
