With 1644 CPUs and a performance benchmark of 5137 GF Rmax, the NZSC would currently rate at number 70 in the list of the world's top 500 supercomputers, says Houston.
These days, Houston's mobile phone rings incessantly, as he deals with clients from all over the region including Japan; the United States, the UK and the Middle East. The clientele list from different time zones partly explains why one of the questions in the FAQ section of the NZSC website is, "Where is New Zealand?"
Another question Houston is often asked is the NZSC's links with Weta Digital, the special effects company founded by director Peter Jackson of The Lord of the Rings (LOTR) trilogy and King Kong.
"We are a partner company," says Houston, who was the chief technology officer for Weta Digital during the filming of the last two movies from the trilogy, as well as Van Helsing and I, Robot, prior to his current post as business development manager at the NZSC in Wellington.
The goal, he says, would be for the NZSC to provide the processing power for Weta as it works on big budget movies, while providing on demand services for other organisations that need massive computing infrastructure to analyse large amounts of data.
Indeed, the NZSC's corporate DNA is inextricably linked with Weta because the concept for the centre was hatched when Houston and his team were working on The Return of the King, the last of the LOTR trilogy.
Weta Digital had by then already built a massive computer processing facility in Wellington. At the end of the filming for the trilogy, Houston says it had 3500 CPUS, of which 1000 were sitting in Wellington at a facility maintained by Telecom.
"We had five computer rooms at that stage and I said to the director [Peter Jackson], 'Look this is not the kind of business that we really want to be in.' Film production is a crazy, cyclical business. The processing power was used three to six months and the rest of the time, it was 'sort of unused'.
"We had all this spare capacity that we knew we weren't going to need until King Kong. So I said, what we really need to do is find a new model that would allow us to take this capacity to the market."
Houston worked with Telecom to develop the concept and formed a partnership that became the New Zealand Supercomputing Centre.
In New Zealand, one of their first users was the Bionformatics Institute in Auckland which is doing research into the evolution of the HIV virus. Their largest customer at the moment is a US manufacturer which uses the supercomputer to look at the thin film layer across fabrication plants and chips.
NZSC provides an on demand rendering system for architecture companies, in partnership with Urban Voyage. "We run the renders overnight and return that information the next day."
The centre is also involved in the beta testing of Microsoft's Computer Cluster Server 2003, as the software giant forays into supercomputing. Just before the yearend, NZSC ran a pilot project to render the images for a Japanese company which is developing a game for the xBox 360.
Houston explains NZSC is "really lucky" to have facilities in Wellington's Courtney Place that are managed by Telecom. Telecom provides the bandwidth and allows NZSC to connect with its customers at 10 gigabits. "We get 24 hour digital and physical security."
Telecom's Gen-i provides the sales channel for NZSC. "Credibility is critical for us here and having a partnership with Gen-i is incredibly valuable for us."
He adds, the "real secret" to their success is a company called Asterisk that was acquired by Gen-i. He describes the company as "a bunch of Linux gurus".
The Auckland-based company is responsible for managing the utilisation of processes and the connectivity of the clients, the operational technical support, configuring the OS, running the jobs and providing the firewalls. So, he explains, the hardware for the centre is in Wellington, but the team providing software support is based in Auckland.
"We are actually a virtual team," says Houston. The only other full-time staff at NZSC is Eric Pilon, service line manager for hosting and storage.
But while Houston is comfortable describing the nuances of grid computing, he vehemently denies, though with a smile, that he was ever a geek. "The funny thing is, I have always wanted to sell computers."
He jumped at the chance to work in a computer company called Idaps one summer in college. He was in seventh form at that time and had been "playing around" with computers. "But not with a huge passion. I hadn't really considered a career in computing."
He told the owners of Idaps he wanted to be in sales. They said he needed to know more about IT. Scott was the first employee Idaps sponsored at Wellington Polytechnic for a certificate in data processing. "It was a new course [then]. We learned how to program in Cobol and we worked on punch cards which is [now] old, old stuff."
Houston stayed for three years at Idaps and moved to London. "I was fortunate to be in London during what they call the 'big bang', when they computerised the London Stock Exchange. If you were a Cobol programmer in those days, you could just about name your salary." It meant a "great lifestyle" for Houston, who would work for six months and then spend the next two to three months travelling all over Europe and South America.
It was also about the same time he developed a "passion" for PCs. "I thought, this was the future. Not mainframes or Cobol." So during the mid-80s, he bought a PC from Amstrad's factory in the Midlands. Amstrad produced one of the first clone PCs, selling them for around a thousand pounds each at a time when the old IBM XT cost up to US$10,000.
Work for nothing
He and his then partner moved back to Wellington in 1987. He wanted to look for a job in the PC industry and went back to his old company which was already named Paxis. He told them he did not want to do Cobol anymore, but was interested in PCs, "This is where the future is going," he told his former boss. But Paxis was not into the PC market and directed him to another company called Metro Computers.
He strolled into the store and said, "Hi, I am Scott Houston. I am a Cobol programmer and I have been in the industry for nearly 10 years. I really got a passion for PCs. I have my own PC. I have taught myself how to use it. I made lots of money in London and I will work for nothing."
Three months later, Houston bought the shares of one of the directors who left the company. The business initially provided technical support for people who bought PCs. "We realised the future is actually in selling them ourselves." So he and his business partner sold clone PCs and imported container loads of PCs from Taiwan. But the margins for PCs had started to deteriorate and they were up against big names like Compaq and IBM.
He felt he could make a difference in another emerging area: Virtual reality. "This was the start of my personal vision in the industry, which is to stay on the leading edge." What got him into this mode was Howard Rhiengold's book Virtual Reality.
After putting down the book, he developed a "real passion" for virtual reality and started reading and researching about it. He left Metro Computers and joined Centron. He stayed with the company for nine months and went back to London to pursue a career in virtual reality and to watch the 1991 World Rugby Cup.
In London, he approached prospective employers with his offer to "work for nothing". The employers there were not as responsive to a Kiwi who has been working in IT for 12 years, was forthright about having no experience in virtual reality, but who described himself as a "fast learner".
"Kiwis kinda admire that and often they give you a chance. In the UK, people would think you are an absolute nutcase, or are suspicious."
But Houston surmises there was also another reason for this reaction. "In those days, virtual reality [coined by Jaron Lainer, just two year earlier, in 1989] was leading edge but it was not commercially viable."
Houston, nevertheless, stayed on in
London "looking for ideas". At that time, the Eastern European market was starting to open up. He and a business partner bought "end of life" PCs - 286s and 386s - and sold these to the fledgling capitalist markets. Their customers did not have cash and paid them in Skoda cars, container loads of coal and thousands of litres of vodka. A Swedish company took all of these products and paid them cash. "It was a bit scary. It was a real bartering system."
He and his business partner also worked out an online broking system. But it was in 1993 when mainstream use of the internet has yet to happen.
Houston used his programming background to develop a fax on demand system. Customers would call an 0900 number and choose from products available from different companies that paid to be listed in the database.
After pressing a number applying to the category of products they were interested in, customers would type their fax numbers and the system would scan the dabatase for available goods and fax the information back to them.
Houston says the system garnered "quite a bit of interest" in the UK but could not be transported to Eastern Europe where the market for the products would be in demand due to the absence of a telecommunications infrastructure for the 0900 number. "It was an idea before its time." But, by then, Houston and his partner decided to return to New Zealand to start a family.
Houston got a job as southern regional channel manager for Compaq, which meant he was in charge of PC sales from Palmerston North to Invercargill. He stayed with Compaq for three years, and then moved to Silicon Graphics, also known as SGI, as an account manager and then latterly as NZ regional manager.
Houston had already established contacts with the National Institute of Water and Atmospheric Research and learned that the agency was interested in buying a Cray supercomputer.
We work to live
But Silicon Graphics did not give him the job that he wanted - working with the research and education sectors, whose portfolio included Cray Computers, then newly-acquired by SGI. The prospect of selling a Cray computer excited him. "If you sell cars, you want to sell a Ferrari. If you sell planes, you want to sell a 747. Well, the Cray is the Ferrari of the computer world."
Instead, he was assigned to the entertain-ment and media sector. He knew little about the sector, but his boss told him, "There is a little company out in Miramar called Weta Digital and they are doing some really good stuff. They have just announced they are doing The Lord of the Rings."
To prepare for his new role, he read the JRR Tolkien tome in three days. "Since then, I have read it a number of times and I have enjoyed it."
Silicon Graphics provided Weta Digital the visual workstations, storage products and servers for production, post-production and visual effects on the LOTR trilogy. At the end of the Fellowship of the Rings, the first movie in the trilogy, Houston joined Weta Digital.
After three years, he decided it was time to move on. "I felt that I had climbed my mountains at Weta Digital and Milton Ngan [then digital operations manager and now Weta CTO] was certainly well and able to take on the responsibilities."
"It was very exciting, very exhilarating but very demanding," he says of those years at Weta. The pace at the Two Towers and the Return of the King "was very full on", and he spent an average of 60 to 80 hours a week at work.
"I had a young family, I hardly saw my kids, I felt I wanted to change my role." He felt the long working hours also contributed to the breakdown of his marriage.
Today at NZSC, he works 40 to 50 hours a week which is "par for the course in this industry". He says: "You have to be quite ruthless in factoring your family into your long-term goals... I don't think you should lose sight of the fact that we work to live, not live to work."
Houston says there are researchers like astrophysicists and theoretical chemists around the world, who could take every CPU at NZSC. "They have got fantastic experiments but they haven't got any money... We came to the realisation we actually need vertical market solutions."
The centre deliberately broadened its application systems for the biotech, gas, geological and nuclear science industries. "These industries had commercial imperatives. These are relatively financially viable so we focus our solutions on the markets that can pay for our services. That does not diminish the fact that we would like to help researchers and there are some good initiatives on that. But the reality is this is a commercial initiative."
The company is developing a system where clients can book for services online. At the moment, the system has been completed for biotech companies, but work is ongoing to provide the same system for other industries. "We developed an interface that allows them to go to the website, pull down a list of the applications, select the data set they want to analyse, state how many processors they want and only pay for the processing time they actually use."
Houston explains the challenge in running a commercial supercomputer centre "is always about utilisation".
With this in mind, Houston is working on developing a distributed computing model using the computer clusters at the NZSC for "commercially sensitive work". But if an organisation will require processing power for generic type of research, the centre plans to tap into computers in schools, universities and businesses across New Zealand.
"The direction I want the supercomputer centre to take is to continue to involve solutions for the vertical markets but also develop a distributed computing model so potentially in the future we could run jobs not just in the supercomputer centre but in homes, schools and universities."
The NZSC could pay for the CPU time that was used. "This is a win-win. We can lower the cost of doing research and we can provide a revenue model or incentive model for the schools and universities." No other supercomputer facility has achieved this. "That's the Holy Grail of the supercomputing market and we are certainly in the race. Next year, we might start to see some solutions evolve around that space."
Houston is also looking at working closely with the government's Advanced Network project which aims to provide an optical network 'backbone' linking research and education institutes throughout the country. The government, he says, can amalgamate all of its research projects and work with the NZSC which will provide a bulk or wholesale rate for the researchers.
"The New Zealand government does not have a centralised tool or mechanism to fund research in high performance computing. And this is the perfect opportunity to establish that."
He is, however, watching how the announcement that the Advanced Research and Education Network project has chosen TelstraClear as its service provider, over NZSC's partner Telecom, would affect this proposed working relationship.
"We can provide a million CPU hours available for research at a fraction of the cost to build it. Our message is, 'Hey, we are here, come and use us.'"
Houston looks forward to the time when spare processing capacity in businesses, universities, home computers and even game consoles can be used "to actually solve real problems that are facing the rest of the world". These could be environmental issues, such as developing a climate modeling in Asia, or searching for water in Africa. He is especially proud New Zealanders can make a difference, without the need to go offshore. "Twice in my career, I had to go overseas to further my career and look for new opportunities.
"But the really exciting thing today is there is absolutely groundbreaking research and cool things happening right here, in Wellington."
Join the CIO New Zealand group on LinkedIn. The group is open to CIOs, IT Directors, COOs, CTOs and senior IT managers.