As companies that conduct their business online have proliferated, hosting of those websites that provide the primary point of contact for transaction of business has become highly mission critical; having these sites go offline or perform poorly not only loses revenue but also detracts from the company's image and loses customer loyalty.

Companies such as betting and online gaming have even more stringent requirements because they need to provide a guaranteed response in near real time. Sports betting events also have the effect of crowding all the business within condensed time windows.

This level of hosting goes far beyond the simple provision of bandwidth and the quality of the server matters not only in terms of its CPU's processing abilities, quantity of RAM and hard disk space but also the durability of its components, the power supply available and the ability of fan/s to disperse the heat generated by a server running 24/7 under possibly very heavy loads.

With an application designed for use internally within a company, one can always put a cap on the maximum number of people that could be using the application. With the internet this number can be unpredictable, or if measurable through registration, can grow large very quickly. The internet is a new operating regime, not only in terms of the security issues it presents but also in the scale of operations, and this requires a new way of thinking when designing applications and the hardware architectures that host them.

The traditional approach is to scale up vertically, increasing the bandwidth, CPU/s' speeds, memory and so forth. There is a limit, however, to how far this can be taken and with so much depending on such a concentration of resources, a failure is nothing less than catastrophic. The answer is to achieve scalability horizontally with a distributed architecture. This architecture not only allows increased scalability, but also creates reliance to faults and failures within the system. What's more, this model is inherently suitable to most operations and services offered over the internet which are in themselves quite simple, but that there is just too many of them.

This situation is akin to how humans organize themselves to accomplish very large workloads; there comes a time when one person, no matter how hardworking and clever will not be able to cope. At that stage the tasks will be split among many people doing exactly the same task and yet coordinating their activities. Imagine if you will, people flooding into the premises of a bank or payment office. Many cashiers wait in booths doing exactly the same job, overseen by managers and perhaps a helper guiding the queues. The architecture and layout of the building hosting these activities, is itself designed to allow a smooth flow of people.

Hosting a distributed architecture is more complex then a traditional centralized system where everything happens in one place. Parallel events need to be coordinated so that they work as a seamless whole and transactional control assumes a critical role. Hardware setup and middleware software need to be designed much in the same way as a purpose-built building, and layers of middle management in the organization would be in place for a human organization. The applications themselves should be aware that they are running in a distributed environment and be able to both benefit and not obstruct this environment.

The key to a successful internet presence stems from both an understanding of the nature of the internet and the tools that are now available to build up this success. The internet has indeed come a long way.

Endeavour has evolved its hosting services starting from its own demanding requirements for hosting its Internet Payment Gateways. Since then, Endeavour has been offering advanced hosting services to serve clients around the world. Real time backups, applications designed for distributed architectures, geographically distributed resources such as databases and fail-safe architectures are at the heart of these services.