Interview with Martin Zinser
Senior Vice President – Head of Section Trading Interfaces at Eurex
Martin Zinser has been using VMS since 1987, when he worked as a student at a nuclear physics lab in Germany. After completing his studies he joined SAS Institute supporting customers using VMS and Unix based “Minicomputers”. Since 1995 he has held various positions in Group Deutsche Boerse. He spent time in operations, design, and development of the high performance trading systems of his employer, a lot of them on VMS, but also lately on Linux.
All the opinions expressed here are his and no conclusion about the position of Group Deutsche Boerse should be drawn out of them.
It has been several months since HP announced their agreement with VMS Software Incorporated (VSI). What were your initial thoughts when you heard the announcement? What are you thinking now, after Boot Camp and having met with VSI and some of the community?
First of all I was relieved that VMS all of a sudden had a future again. Obviously the announcement by hp to officially cease development of VMS had maybe not been a surprise, but a big disappointment never the less. With the initial announcement about VMS Software Inc. I felt very confident about support for i4 (since rumor had it, that this was mostly done already) and the Kittson based Itanium servers. It was nice to see x86 on the roadmap, because that is very obviously the way forward. But at this point in time I it still seemed pretty vague and more like a marketing item to be ticked off. Having met the VSI crew at the Boot Camp and hearing about the steps they have already mapped out, the meetings with Oracle engineering to discuss the tool chain needed to get their products ported, the project seems much more real now and I feel confident that they will be able to get us there.
Based on what you know what would you count as VSI’s key strengths?
Focus and enthusiasm for VMS. This is not just a job for them, but something they really care about and want to see succeed. This is paired with long experience in the industry both on engineering and businesses side to give them a better chance to succeed than other start-ups have.
What weakness have they inherited and what weaknesses are inherent in their position?
One of the main problems for VSI is that their installed base (Enterprise customers with needs for enhanced security and reliability) is rather conservative. An as yet unproven start-up is typically not what these customers are likely to buy from. So VSI has a two-fold challenge ahead of them. On the one hand they need to deliver updates to VMS in-time and in good quality, i.e. V8.4-1H1 needs to be out best in in the first half of 2015, but certainly not long after. On the other hand they need to proof that their business model is working out and that the company is financially sound and will be around in the long run.
Where do you see the great opportunity for OpenVMS growth, if at all?
Once VMS becomes available on smaller form factor systems (e.g. x86) there are lots of areas where growth is possible. One obvious example is ATMs. Each transaction there has a considerable monetary value, but security in this area currently certainly is not as good as it should be. Having VMS systems already at this tier of the application, will allow building much more secure and robust solutions. Going from there other applications in the electronic payment realm would be the next natural target.
What are the biggest risks facing VSI and the user community?
A lack of momentum in the user base going forward. Due to the neglect of VMS by hp in the last couple of years, culminating in the announcement to stop development, many customers either already have made plans or even started projects to move to other platforms. Many other installations moved into maintenance only mode. Given the structure of the agreement between hp and VSI, VSI needs customers to actively upgrade to new versions of VMS (typically together with new servers). So first existing customers need to get energized and again look at their infrastructure and applications and start investing in them. In the end this is what will allow VSI to thrive and deliver the future of VMS we are looking for.
How do you react to the statement, “The oldest thing in most OpenVMS shops is the development process, not the software or hardware?”
Don’t know, it seems some shops out there have pretty old hardware, if there are still VAX systems in use 🙂 As far as the process is concerned, it is really more a question of the culture in a particular company that determines what is used and how successful it is. There can be no process, waterfall approaches or agile methodologies and all the shades in between. All of these are possible if you are a VMS shop.
Cell phone makers realized long ago that they needed happy productive developers in order to win the market. Do you think the quality of the development experience has a similar affect in the enterprise space?
Yes, I do and this has been discussed often in the VMS community. One of the steps in the wrong direction (still made by digital) was to abandon the “low end” market, including developer workstations. Developers who are comfortable to work with certain tools and environments are much more likely to recommend a solution based on what they know and trust, when asked for a design for a new service. If there is a disconnect between the developers environment and the “enterprise” solution, developers will push for their use of the environment they use and trust at the enterprise level. This has been the case during the initial adoption of VMS, and certainly has played a big role in the success of Sun and Microsoft Windows based systems later on.
What tools and capabilities would make the biggest difference in the future success of OpenVMS and VSI?
First of all (and sadly enough) the compiler infrastructure needs to be brought up to current standards again. DEC compilers used to be really good in the past, but also here some neglect is showing. This is a pre-requisite to bring VMS also up to speed (and included in the upstream) of important Open Source projects. No system is an island anymore; applications need to communicate with each other. This not only means basic network connectivity (TCP/IP services being another area that needs a thorough overhaul), but also messaging and data formats. In many of these areas Open Source products are the de-facto standards by now and there is just no way around to support and hopefully strengthen them with the experience and knowledge brought to the table by the VMS community.
What role will a flexible OpenVMS IDE, such as NXTware Remote, play in the future success of OpenVMS?
It is one way to ease the transition of new developers onto the VMS platform. Eclipse with its plugin architecture is pretty much ubiquitous in development environments currently. For many younger developers it is the only style of development tools they are used to, since they never have been seriously exposed to command line environments (be it Unix like systems or VMS). So finding the same interface on VMS will allow them to concentrate on the advantages and conceptual changes of the platform, while avoiding distraction and annoyance by being forced to get to learn to use a new development environment.
What advice would you give companies about their use of OpenVMS, especially those on the fence about the future of OpenVMS in their environment?
Get to know the team at VSI. They are approachable and open for discussion. You will learn that they know what needs to be done, they are aware of the challenges, but they also have a plan how to tackle them. This work is their passion and they are not going to give up on it. They do want to do, what’s right for us as customers, and this is more than can be said about many other organizations.
Some say perception is reality. How can developers change how corporate management perceives OpenVMS?
Corporate management really does not care that much about the technical merits of one operating system versus another. What they do care about is vendor risk and business solutions. Vendor risk has been high in the last couple of years for VMS with the strategic direction of hp. Obviously VSI is a big step in the right direction there, but they still need to proof themselves. As far as business solutions are concerned, developers need to blend the unique strengths of VMS together with the rich Open Source ecosystem out there to create business value. Speed, quality and cost can not be improved in leaps and bounds all at the same time, but there is certainly potential for VMS developers to provide superior value on all of them if we keep an open mind.
What are your greatest hopes for the platform and OpenVMS community?
That we will build and witness a true renaissance of VMS. Many concepts and features in VMS are still superior to the state of the art in other commonly used operating environments. But time has not stopped, and there are certainly things both on a technical and community level we can learn from the success of Linux. If we manage to reignite the passion of the existing base and can draw more young people into the community and foster a spirit where everyone is taking seriously and valued for the contribution they can bring to the table, we can see VMS spread its wings like a true phoenix.