“We are growing incredibly rapidly and scaling our development teams is the biggest challenge,” says Simon Roach, chief technology officer at KEMP Technologies.
US load balancing company KEMP Technologies is to create 50 new jobs in Limerick over the next three years, bringing its current Irish workforce to 80.
The company, which has seven offices worldwide, with Limerick heading up the EMEA region, is a developer of load balancers that enable all types of SMEs and global groups to optimise their e-commerce and online traffic capabilities, such as online banking, email, web and cloud-based applications.
The company also intends to invest even more resources to accommodate its rapid expansion in the Limerick’s National Technology Park.
What are the big trends and challenges in your area?
I’m based in Munich. We obviously have the R&D facility in Limerick and I spend a fair bit of time there and in an office in Santa Clara and we have a core dev team here in Munich.
There would be 35 people in the core R&D team and the worldwide support team report to me additionally. We have an Irish guy who runs the worldwide support team who is based in Long Island, New York, and we located him from Limerick to Manhattan, which must have been a culture shock.
We are spread across locations in Long Island, Limerick and Singapore – we follow the sun.
We are growing incredibly rapidly and obviously scaling the development teams is one of the big challenges we’ve been working on for the last 24 months and Limerick has been very, very good for us so far.
What factors led to KEMP deciding to make Limerick the company’s EMEA headquarters?
The founder of KEMP had chosen Limerick back in 2009 and obviously we wanted to have a European dev base and an EMEA headquarters and Ireland was the lowest-hanging fruit in the sense that the language was the same. That was a big factor from a location point of view. The UK would have been the other option, but the close vicinity of the universities, such as Limerick Institute of Technology (LIT) and University of Limerick (UL), whom we work with, was a big deciding point.
The relationship and the way that LIT, in particular, has engaged with us has just been unbelievable. We’ve taken on a large number of graduates – five or six students in their final year doing projects with us – this is the fourth year’s intake we’ve had in and most of them have joined KEMP once they have graduated. They work in software engineering, software development and our customer support teams.
We have been getting first pick of the next batch of students from the network and computer science courses based on good experiences that people had.
One of the things we do quite well is we are very respectful to undergraduates when they come into the company and we treat their bachelor’s projects as if they were KEMP internal projects so they treat it as a professional project. We track it through all of our processes and give them the time and the focus that it deserves and that certainly has persuaded a number of guys to come back.
A couple have got first-class honours degrees based on their good projects. That has worked very well.
It is working less well on Long Island and we are only now starting to engage more with some of the colleges and universities there and using the model that we developed in Limerick as a kind of a guiding light.
We want to expand this to Cork, as well, and engage with Cork Institute of Technology (CIT) and University College Cork (UCC).
In terms of acquiring fresh talent, it has been great. Obviously, acquiring more experience in the short term is starting to become quite tricky. The past 18 months has become very difficult. It is a worldwide phenomenon, it just feels particularly difficult in Ireland because of the size of the population.
How complex is the infrastructure, are you taking steps to simplify it?
It’s complex but it comes down to skills ultimately. First of all, getting the kind of people we need fresh and trained up is very, very difficult. If I can give our QA processes and the way we go about quality assurance at KEMP you will find there are not many methodologies out there that suit us and not many companies do it probably as well as we strive to. There aren’t many people with finished skills.
If you look at it from a development point of view, we are looking at people with Linux skills, people who have worked on the Linux kernel, people who worked with protocol stacks, TCP/IP, people who’ve worked on the support side with the kind of enterprise workloads we support, such as Lync, Oracle, SAP, Exchange, so depending on where we bring people into the organisation there’s a varying skill set that is required.
What are the main points of your company’s development strategy?
We have a strategy in the R&D department which is built on three pillars and we call them our technology core values, like companies have core values on behaviour and so forth.
Technology core values sounds corny but it works for us, it gets everyone on the same page.
The three areas are: platform ubiquity, interoperability and API economy, and the third area is platform richness.
The idea of that within the area of application delivery is that it is not just load balancing, but the broader area of application delivery.
On the ubiquity value we are looking at porting software because primarily we are a software company. To the outside world we are a company that provides gold boxes, appliances and virtualised software, but internally we’ve always been a software company.
We build an operating system that is standardised, we put that out on standard x86 hardware, which in some cases is accelerating, but fundamentally the way that the company ticks is the way that most software companies would tick. What we are trying to do with our operating system KEMP OS, which is a Linux-based operating system, is to make that available on as many platforms and as many creatures as possible.
How diverse is the range of platforms you are addressing?
Today we provide our ADCs on gold appliances, which was the traditional business. We are also starting to provide them as bare metal – you can buy a bare metal OS from KEMP that will run on C platforms, HP, Oracle, or a Solaris-based system, it will run on a number of diverse systems and then of course we have running the OS on the cloud-based operating systems and hypervisors, so there is a very broad-based operating system component here that we need to get in place.
Then we have the whole API economy and interoperability stuff where we starting to find ways of better embedding our technology to other environments, be it orchestrators with OpenStack and System Centre, SDN and NFV, as well.
SDN is an area we are looking at for people at the moment. We hired a young guy from LIT and he’s been with us for six months on the support side and he’s a very good programmer with some great networking skills and we are in the process of building an SDN integration test lab in Limerick and that is one of the things he will be working on.
But to complete the three values – platform expansion – we want to expand the platform into other areas, tick more security boxes. We launched a security pack last year and we are in the process now of bringing a WAF application firewall and tailoring our WAF approach and making it application centric.
That is one of the strategies we always followed at KEMP as opposed to competitors – putting big pieces of IT infrastructure in place that can be used by everybody for all sorts of processes.
It has been said that if software is eating the world, APIs are eating software. What challenge does this create for KEMP?
You’ve hit it on the head – our ubiquity is about supporting as many platforms as possible and we support something like 14 or 15 platforms at the moment so whenever we make a release on an eight-week cycle we try to maintain this religiously – this includes 13 or 14 OS platforms every time we release something, two cloud platforms, a large number of hypervisors and bare metal platforms for Dell, Cisco, HP and Oracle.
Every time we go through a release cycle one of the big challenges is making sure our software operates on all those platforms. It was clear from the outset that we had to create a highly automated test lab so from day one we put a lot of effort into process automation and test automation, and we’ve got literally hundreds of thousands of test cases that have been coded and programmed.
It is a continuous challenge to support more platforms and broaden and deepen the test cases, in particular, adding functionality and the APIs is the big challenge.
We are a load balancer, in the middle – above us are orchestrators and below us platforms like SDN controllers, and just coming to terms with the number of these APIs is a challenge. So supporting the OpenStack, being involved in the committees, being a member of OpenStack, the orchestration side is challenging. We have our own Restful API and try to adapt that – but just tracking all of this a challenge. A challenge we relish all the same.