Cloud Computing: Server Hosting with FlexiScale

Background

A month or so back a client came to us with a requirement for a dedicated server that would host an industry-specific application running on Windows Terminal Server. It was quite a simple application and the company concerned had the in-house server capability to deliver it. Yet there were one or two snags:- the client themselves were not a 24x7 operation (the application had to be available round-the-clock to support their worldwide customer base) and they did not have fast, redundant links to the Internet. Hosting the service internally was just not an option.

The standard alternative is usually to look at co-locating a server within a data centre somewhere. After all, 24x7 operation and huge amounts of bandwidth are pretty much bread and butter to these guys. However, the size of the application and the initial roll-out plan didn't really warrant purchasing dedicated hardware. This got us into looking at cloud-based server providers who could provide us with an 'on-demand' service using a standard virtualisation platform.

FlexiScale

Which brings us on to our chosen solution:- FlexiScale. An on-demand cloud computing provider that is a subsidiary of the UK web hosting company XCalibre Communications. The key features of their service were an ideal fit for our requirements, namely:

Servers can be deployed instantly.
They can be scaled with demand.
Multiple Operating Systems are supported (including Windows Servers)
Pricing is pay-as-you-go with no initial investment or long-term tie-ins.

Deploying the Server

Getting the server up and running proved to be a snap. Once you have signed-up via the FlexiScale web site you just need to login to the control panel, add some initial credit and then you can start provisioning your server:

A view of the FlexiScale Control Panel

There are a decent number of server images available including CentOS, Debian, Ubuntu, Windows Server 2003 & Windows Server 2008. Decide on the amount of memory and disk space you need and your server will then be provisioned in real-time and be ready to use immediately.

In terms of cost, the price per-hour is broadly a function of how much RAM you require with some smaller additions for a firewall, disk space and monthly bandwith utilised. Windows Server operating systems also incur an additional charge to cover the cost of licensing.

Configuring the Server

Once our Windows 2003 Server was up and running it was a quick and painless process to get it ready for use. The server itself is managed using a Remote Desktop Connection which allowed us to connect from our office and install the specialised application inside Terminal Services.

After testing the software worked ok, the next step we wanted to do was to setup a firewall to protect the server on the Internet. Again the FlexiScale control panel makes this quite a simple affair and allows you to setup the necessary access rules for inbound and outbound traffic. In this case we just reject all traffic except for known hosts connecting via Terminal Service port 3389.

Our last step was to install a monitoring agent on the server that would alert us to any unexpected downtime or performance issues. We're quite big fans of the GFI MAX service (formerly HoundDog) and find that it makes looking after servers far, far easier than it would otherwise be. Doubly so in this case as we have no physical access to the server in the data centre.

Going Live!

The solution went live to our client's customer base last week. As of yet, we've not experienced any downtime and the customer is entirely happy with the delivered solution.

From our perspective the service was quick and easy to setup and is simple to manage. It is resilient and secure and will scale as demand grows. Our client had no initial outlay or investment and can stop the service at any time. What's not to like about all that?

So next time you need a server, be it short or long-term give 'on-demand' cloud computing a try...

Bookmark

Trackback(0)

TrackBack URI for this entry