In the beginning . . . .
Back in late 2000 our operating system and networking development teams, based in Munich, were faced with the problem of scaling data compression compute power for a product innovation we had created within the area of Intelligent Networks (IN). IN is a standard ITU network architecture used to deliver value-added services in both mobile and fixed telecom networks. IN was an attempt in the 90’s to remove service control call logic from the service switching points which were predominately ASIC (application specific integrated circuit) based and place this logic in a centralised layer of flexible service logic which was software driven.
After some research our architecture team came to the conclusion that the solution to our scaling problem was something called an IP Load Balancer. However, looking at the products available at the time, they seemed both overly expensive, driving the total cost of our solution up considerably, making it much too complicated to operate and thus driving up the total cost of ownership.
Being experts at implementing service control switching technology on standardised UNIX server platforms, and with many years of experience implementing IP protocol stacks, we felt we had a sound understanding of the principles of performance networking and high availability, and so our rapid prototyping team set about building an appliance of our own.
Portability, Value for Money and Simplicity of Use
Key to the solution was that it had to be Robust, Portable, Provide good Value for the Money and be Simple to Use. Consequently, the system was based on the Linux Operating System running on commodity-off- the-shelf (COTS) x86 server hardware. Despite Linux being in its early days as an appliance platform and thus a contentious choice to make, our skilled developers were adamant that it was the optimal choice, and with “some” tweaking, was fit for its purpose.
Much of the load balancing logic was embedded into the kernel and the hard disk was replaced by flash memory. SSL traffic was off-loaded to a dedicated ASIC supported by high performance drivers that we developed in house. An intuitive graphical user interface was created that distilled all necessary functions and operations down to a few mouse clicks.
The engineering disciplines and productisation skills we had cut our teeth on, working for leading global server vendors of the day and some of the world’s largest telecoms ensured that the product was “as solid as a rock”.
Whilst in the early years the product was not a roaring success, it did gain a steady and loyal following of users that appreciated both the reliability and ease with which the product could be integrated into their networks. Furthermore, operators of line of business networks loved KEMP’s application-centric approach.
How we have come along . . . .
From the early days of supporting only a single LoadMaster platform with single-site development capabilities in Munich, today KEMP with its headquarters in New York City, supports an array of branded appliances, bare metal platforms (such as Cisco and Dell), both branded (Microsoft, VMware, Oracle, etc.) and open source hypervisors, as well as cloud platforms such as Microsoft Azure and Amazon’s AWS.
As a global company with tens of thousands of licensed products and product development facilities in three geographic regions (US, Ireland and Germany), we have had to “up our game” considerably in terms of innovation, productisation, execution and customer care.
Agility is paramount
Agility is paramount with full KEMP operating system releases being made on an 8 week cycle. In doing so, an obsessive sense of quality remains at our core, with one quality assurance engineer for each and every developer. Continuous software integration combined with test automation systems run hundreds of thousands of test cases around the clock ensuring both quality and scale.
Our highly committed customer service engineers in the Americas, EMEA and APAC regions are deep application workload specific subject matter experts. They serve our customers around the clock with support queries and professional services.
An obsessive ability to focus and reflect on the customer benefits of new “game-changing” technologies such as SDN (Software Defined Networks), NFV (Network Function Virtualization), Service Chaining and Cloud Orchestration remains deeply ingrained in this dynamic environment. One thing however has not changed to this day . . . our commitment to Robustness, Portability, Value for Money and Simplicity of Use have remained core values in guiding the evolution and creation of KEMP products into the future.