Developers | Claranet UK


Why independent software vendors should consider a SaaS model managed by a managed services provider

Software-as-a-service, the process of delivering and accessing applications through the web has become the defacto method of software deployment.

End users want to access software on any device, anytime and anywhere. Independent software vendors want to be able to move toward a cloud based service model to meet customer demand, stay competitive and introduce recurring revenue streams.

The problem then is working out who should support the delivery of your software. At the outset there appears to be two options: either manage it in-house – requiring extra resource and expertise you may not possess– or work with a hosting or managed services provider.

In reality there may however be three options. Hosting providers only cover one part of the picture, the end platform, what they are not able to do is manage the application experience from the data centre to the end user’s device. This leads to a grey area where nobody is taking overall control for the application’s successful delivery. If there is an issue where performance is impacted or downtime occurs, this can turn into a finger pointing exercise between the host, the network provider, end users, and the software provider, as they seek to apportion blame on each other.

Managed services providers can provide the end-to-end infrastructure and management expertise to most effectively deliver your application. However, not all managed services providers are equal, and working with less evolved providers who don’t take the time to understand your business goals can lead to damage to your brand and revenue.

A customer's perspective: Journeying to the cloud - migrating away from a legacy platform

This blog is used with permission from Sift, the digital agency, who are a customer of Claranet's. The original blog can be found here.

A few years ago a venture capitalist told me the secret of success when growing a technology business was to sell the company within a single technology cycle. His reasoning was that you really didn’t want the hassle and risks of transitioning your platform from one paradigm to the next. This post is for those of us with longer term aspirations.

In our case, the project started in 2007 and culminated six years later with Sift’s CIO Chris Wood switching off our servers at Verizon just before Xmas last year (as evidenced by the empty racks and redundant servers below), consolidating everything with Claranet and Amazon.

Developer? Is managed application hosting right for you?

Claranet has offered managed application hosting for many years, but when you look out into the hosting market I am constantly surprised by other hosting companies running for the hills at the very thought of taking over the control and ownership of a customer’s live environment, and offering an SLA based on anything above the OS.

Why are most hosting companies scared of taking control above the OS?

It's clearly no mean feat, you simply can't take over a customer’s application and not expect issues. You have to learn how the customers application works, the issues it faces, and devise a contingency plan just in case something goes wrong. Organisations need to invest in processes such as ITIL which is a best practice framework for the provision of quality IT services, build the processes and documentation, and ensure they have sufficient qualified engineers to be able to support an application and all the infrastructure around it 24/7/365.

Why should developers care about managed application hosting?

It's simple, the developer’s core focus is usually based around delivering great code. However, we typically find that they somehow manage to also get burdened with having to run the hosting environment as a side-line. This generally happens because they possess the technical ability and understand their app, but running infrastructure is not what they have studied for years and trained for, and when problems happen this diverts their attention from core business projects.

By letting Claranet manage the application, developers can concentrate on spending more time delivering code, meeting deadlines, and adding more value to the strategic projects their organisation demands.

How virtualisation can help developers part 2 (testing and deployment)

This blog follows on from How virtualisation can help developers part 1 (development) and explains how virtualisation can continue to benefit developers during the testing and deployment phase of a development life cycle.

• Testing (Functional) – Within the development life cycle, there is always a form of functional testing, evaluating whether an application functions as it should. This may be narrow in scope, i.e. testing the payment gateway works as expected, and may require an environment specific to the modules/code being tested. In a virtual environment an application environment can be spun up, tested and then destroyed. This approach is beneficial to businesses because it reduces costs (you only pay for what you use) while also increasing the speed and agility within the development cycle.

• Testing (Load) – Prior to making the application live you may want to know how it scales and how it will react under load. A virtual environment like Claranet’s Virtual Data Centre can be scaled easily and quickly, with its drag and drop interface and through the use of templates. Again, a load testing environment can be spun up with minimal effort and then destroyed after use.

• Testing (User Acceptance Testing) – Prior to any major code release the end customer will want to confirm for themselves that the application is fit for purpose. For the reasons given above, virtualisation and a platform such as Virtual Data Centre will easily meet the requirements of User Acceptance Testing.

• Preparation for production hosting – During the testing phase you will probably create templates to increase the speed of deployment. These templates can be recycled for production as they are ready built images. Using the sizing data obtained from the load testing, you can now quickly deploy a solution that will meet the client’s expected peak requirements. If the client’s application has anticipated peaks in demand - for example a television program that airs in January until June every year where the peaks are massive, but outside of that time the site processes a low number of hits per second - you can scale resource allocation accordingly using the drag and drop interface in the Virtual Data Centre.

Subscribe to RSS - Developers