Cloud Machine Manager

Cloud Machine Manager is a web-based cost-saving tool developed by Blueberry for Amazon EC2 cloud users.

Amazon Web Services allow companies to purchase a virtual machine in the cloud on a ‘pay as you go’ basis, which is great and can be a big cost saver in itself. However, just like other things in the office, many people forget to switch off their Amazon machines at the end of the working day. When paying by the hour, this can soon add up to a significant waste of money. We experienced this ourselves and decided to do something about it – so we created CMM.

CMM automatically turn servers on and off.  When a user stops working, it notices the server isn’t being used, and turns it off.  When the user returns it can turn it on again.

Blueberry designed CMM as a cost-saving tool for small to medium sized enterprises who are working with virtual machines but not 24-hours a day, every day of the week.

For intermittent EC2 users:

• CMM can instantly turn servers on or off on demand.
• CMM can be configured to switch a virtual machine off outside of office hours and back on again ready for the next day’s work.
• CMM can be configured so as to allow an employee to start a virtual machine on request – a job usually requiring a call to technical support and unnecessary delays.
• CMM can watch the usage levels of the machine, recognise when it’s not being used, and switch the machine off.

We chose to implement the core machine management functionality in Linux to allow it to operate on smaller servers.  This also allows us to deploy the core system within customer networks without incurring Windows licencing costs. Additionally, it meant lower cost overheads for Blueberry too.

The system was designed from the start with scalability in mind. A front-end server handles all user interactions; the back end is responsible for all scheduling and Amazon API interactions. This means front end servers can be added in order to service many more customers. The system architecture supports both a multi-tenanted public system, and smaller private deployments within larger customers’ private Amazon infrastructure.

It utilises modern JavaScript web frameworks in order to provide a rich user experience.

CMM also offers cross-platform support with its own standalone mobile applications and installable Desktop Starter App software.

CMM needed to be easy-to-use in order to encourage a wide cross-section of users to try it, in particular SMEs and smaller companies who were reluctant to use – or were inexperienced with – the complicated AWS console for server management and scheduling.

Initially, we developed a smart scheduler to turn servers off and back on again. However the real challenge was to develop the on-demand feature, because this delivers the biggest saving. Importantly, there didn’t seem to be a tool in the Cloud marketplace that could implement real-time on-demand control of servers – which makes CMM unique as a SaaS Cloud product!

One of the difficulties in developing true on-demand capability was figuring out how to intercept web requests for automatically starting the server. Having achieved this, one of CMM’s clever features is being able to start up a server that hosts a website simply by browsing to the site using a URL.

We’ve created a tool that provides real-time on-demand scheduling and automation of Amazon EC2 servers to eliminate pay-as-you-go charges for unused server time. CMM allows users to quickly and simply turn AWS EC2 servers on or off automatically, based on actual usage. Set-up can be completed in a few minutes – you just need to be prepared to move your test domains to AWS Route 53 cloud DNS service and EC2 servers.Cloud Machine Manager:

• Automatically detects and turns off unused servers.
• Switches servers on and off based on a schedule.
• Restarts servers on-demand.
• Scales with demand as you add or remove devices.
• Powerful UI for setting schedules and thresholds.
• Multi-platform technology for potential savings across an Amazon server estate.
• Ability to start servers remotely with an app for Windows PCs, and a mobile starter app for smartphones (both iOS and Android devices).

We estimate that a company using CMM and only running servers between the hours of 9:00 am – 5:30 pm could save as much as 76% in costs, whilst a company using CMM and only running test servers for 10 hours at intermittent times during the week could save more than 90% on their server costs.