Gandi’s mantra for cloud computing
As the cloud computing space grows up faster than a bamboo plant on steroids, so arguably does the sophistication of the management tools that aim to serve it. But with these tools in mind, isn’t a good proportion of what we read about the cloud right now overly focused on the customer delivery element of the cloud proposition? If you feel this is the case, then perhaps we should question ourselves at this point.
Last week my inbox felt a nudge or two from Gandi, a company that produces management software to, it says, ‘automatically’ manage peaks and troughs in virtual server usage. The product is called Watchdog and it claims to provide a monitoring service that allows customers to set certain thresholds on key server performance metrics such as CPU usage, load average, or network traffic. If a threshold is exceeded, the tool will either notify the customer through email or instant message, or will take a pre-determined action such as adding server resources.
Whether or not the company’s French founders recoginse that using a skewed spelling of the name of the pre-eminent political and spiritual leader of Indian independence as a company name is slightly strange I am unable to say. Whether they further miss the relevance of naming their product after the BBC’s highly popular consumer program, I’m also not commenting on.
What I will comment on though is whether this is the type of system-admin level cloud development that we should hear more of if we are to comprehend more of the mechanics that drive cloud space as a whole. Spikes and peaks on virtual servers that necessitate restarts cost everybody money. I just wish I could find a system administrator to comment on this, but they rarely act as spokespeople do they?
Are you in this field? Have you been lying awake at night waiting for out-of-the-box monitoring to manage your virtual stack? I guess perhaps not. But Gandi.net COO Joe White reckons that a high proportion of data centre servers – even in virtualised environments – still run at as little as 10-12% utilisation because data-centre managers allow a large fail-safe for peaks in load.
Could Gandi also be serving up good news for developers? If you start out on a project, you don’t typically know the full extent of the resources you’re going to need. But of course with a virtualised backbone on the cloud, you can create and trash virtual servers and automatically expand or contract the amount of resource you need. If it were a well-managed backbone then you wouldn’t mind putting your application development reliance upon it more heavily. Surely?
As the great Mahatma would have said himself: There is more to life than simply increasing its speed.