Interviews by Stephen Ibaraki, FCIPS, I.S.P., MVP, DF/NPA, CNP
Trevor Eddolls - Internationally Recognized Technology Thought Leader, Author, Editor, Consultant; CEO iTech-Ed Ltd; Editor Xephon's Update Publications
This week, Stephen Ibaraki, FCIPS, I.S.P., DF/NPA, MVP, CNP has an exclusive interview with Trevor Eddolls.
Trevor Eddolls is CEO of iTech-Ed Ltd, a UK-based company he founded at the start of 2004. Trevor has written nine mainframe and communications-based articles for NASPA's (http://www.naspa.com) Technical Support journal and has been responsible for the Glossary section of The Arcati Mainframe Yearbook (http://www.arcati.com/newyearbook07) for the past three years. He has also published three books on mainframe computing.
Trevor is responsible for the creation and editing each month of Xephon's (http://www.xephonusa.com) Update publications. These are written mainly by systems programmers and other technical people for systems programmers, etc. They are full of code and up-to-date information about mainframe-related topics such as CICS, DB2, z/OS, RACF, TCP, and SNA. There are also monthly Update publications for WebSphere and AIX. Trevor has a regular weekly blog, which can be found at www.mainframeweekly.com and also at http://blogs.ittoolbox.com/hardware/xephon. Here he discusses developments in the mainframe world and sometimes computing in general.
Trevor has written a number of documents that have been used both internally and externally by major international companies in the mainframe and communications sphere. He has also helped with the design and development of Web sites for a couple of companies.
When not at work, Trevor is a governor at a local secondary school and also chairman of a charitable trust at the same school. He enjoys spending time with his family and playing with computers. For recreation he likes playing racket ball, swimming, and juggling, and can often be heard playing his guitar.
The latest blog on the interview can be found in the IT Managers Connection (IMC) forum where you can provide your comments in an interactive dialogue.
Opening Comment: Trevor, your record of accomplishment provides a unique perspective for our readers. With your very busy schedule, we thank you for taking the time to share your deep insights, experiences, and wisdom with our audience.
A: No problem - I am always pleased to help.
Q1: Please profile your current roles and activities.
A: The major part of any month is spent talking to contributors and getting articles for publication in Xephon's various Update journals. The rest of my time is spent on a mixture of activities. It is important to stay up-to-date with what's happening, so, for example, recently I was in Mainz Germany with IBM. I write a weekly blog and regularly contribute mainframe or networking-related articles to NASPA's Technical Support magazine. I also undertake some consultancy work, and prepare documentation for use by companies in the mainframe space. This past year I have designed and implemented a Web site for a local company. So I tend to be quite busy with a variety of tasks.
Q2: In your current roles, what are your five greatest challenges and how are you managing them?
There never seems to be enough time to get everything done - or if everything gets done, there never seems to be any time left to do anything else! Now, let's be clear, I enjoy being busy. I don't like to wake up in the morning without being able to think of a list of things to do (a short list is preferable!). However, I often find that I am going back to work in the evening and spending a few hours at the weekend catching up or completing non-core tasks that just need to be done. There is no solution to the problem. It's not possible to take on more staff and half the burden because the work often is linear and therefore the second person would spend a lot of the day standing idle. Management is best achieved by creating lists of tasks that need to be done, arranging them in the most efficient order, and ticking them off as they are completed. It is important to use rewards as tasks are achieved - nothing too grand, usually eating a favourite food or making a pot of tea. It's also important to realistically estimate how long a task will take so that not too much is attempted in day. And lastly, I now try to schedule a swimming session at the local pool. That way I do take a break and get some exercise.
- Finding qualified people:
It's very easy to find people who have an opinion on SOA, virtualization, "green" machine rooms, Vista, etc, but it can be very hard to find people who actually know what they are talking about. Each month, I need about twenty people to contribute articles to the Update publications, and sometimes it can be quite hard to find people who are working with CICS or DB2 (or whatever) who not only understand how the systems work, but are also able to explain it in a written article. There is no problem finding people who will enthuse about a new technology, or who have used the technology, but finding someone who not only understands the technology but also has the time to write about it is very tricky. I try to manage the situation by keeping long lists of people that I know have written articles in the past or answered questions on message boards, I then regularly invite them to contribute new articles.
Running a small business depends on people paying promptly. My company employs people who like to be paid regularly for their work (not surprisingly). There are a number of companies that are very prompt payers - and that is much appreciated. However, sadly, there are others who pay on 60 days or even 90 days, which can cause some cash flow problems. Every company has its own procedures, and there is no way of getting them to change. I thought I'd air the problem of slow payment to small companies and the problems it can create in this forum. I manage the situation by keeping a very close eye on my bank account and ensuring that we are doing enough business each month to cover this delayed payment.
I have recently upgraded my XP laptop to a very nice HP Pavilion that came with Vista. Why is that a challenge? The answer is because it has caused me huge problems.
To a large extent my company is a publishing company - and traditionally publishing companies have used Macs. I had software installed on my XP laptop that allowed me to connect to Macs over a LAN and share files etc. The software is called PC MacLAN and is now owned by CA. CA has no plans to release a Vista version of the software. As a consequence, my new laptop can't talk to the Macs. Secondly, I have an HP Business Inkjet 1100 printer connected to my XP machine. It was meant to be a top-of-the range printer when I bought it three years ago. Are the drivers for this printer included in Vista? No. Are the drivers available for download from the HP Web site? No. So now I'm left without a printer as well. I used a Zip drive from Iomega that was connected to a parallel port on my XP machine for back-ups. Now, obviously (perhaps) my new laptop doesn't have a parallel port. But even if I bought a USB-to-parallel cable I'd find that the Iomega Web site doesn't provide a driver for Vista. I like the Windows + alt key combination. I find "Search" to be simply bizarre. That's why Vista has been a major challenge.
In terms of managing the situation, I work on my new computer (17" screen is very good), and print etc from my old one. I'm hoping HP will make Vista drivers available for the printer or else I will have to get a new printer. And if anyone knows how I can connect Vista to a Mac network, please let me know.
- Keeping up-to-date:
Change in the world of computing is the natural state. I find it important to be able to identify new trends in computing and keep an eye on them as they develop. I try to do this in three ways: I read lots of articles and White Papers, I get sent lots of press releases, and I go to meet vendors (although I avoid anything too "selly"). Even so, with so much going on, some things still slip under the radar. Often a conversation at lunch can be more fruitful in terms of direction of development than a 30 minute PowerPoint presentation!
Q3: What are the five critical issues facing organizations today and how can they be addressed?
A: The five critical issues facing organizations today are staff, training, compliance, government, and flexibility.
Staff: Looking at mainframe sites, it is at once obvious that the staff who understand COBOL and how the system was set up are getting older. In the past, younger people would be employed with a knowledge of COBOL and an appreciation of how CICS and IMS work. Now this is not true. Universities do not teach COBOL, focusing on C++, Java, and .Net. So, the average age of mainframe staff is now 45-50 years old. This is a ticking bomb and in ten years time, a lot of COBOL programmers could be earning a lot of money as consultants, spending a couple days each week with mainframe companies that have no choice but to buy their expertise. If the oft-quoted figure of 80% of data residing on mainframes is true, then the importance of staff who understand how the applications work cannot be stressed too highly. In addition, if SOA applications really do take off, then all mainframe sites will be in the position of having many customer-facing applications that rely at some point on mainframe-resident components. In the event of poor performance (or no performance) someone has got to understand what is happening on the mainframe.
Training: Training leads on from my previous critical issue. Training is an insidious expense. Not only does the company have to pay for the training course, it might also have to pay for meals and accommodation while the person is being trained, and it has to pay for someone else to do that person's job while he is away on the course. It makes the total cost of a training course seem prohibitively expensive. Plus, many companies worry that having trained their staff, it makes them more poachable by head-hunting companies - and they have to start the expense of training all over again. But, and this is a big but, unless people are trained in CICS, DB2, IMS, etc soon, there won't be anyone out there with any real experience able to run these training courses. And, with the growth of SOA, it becomes even more important that the structure of a composite application is clearly understood so that any problems can be identified and resolved. Training is a very necessary expense.
Compliance: Compliance is another big issue. It adds to the cost of doing a job. It becomes an even bigger problem for companies operating in more than one geographical area where compliance might have to meet two or more stringent sets of regulation. Having said that, companies must ensure that they are compliant with Sarbanes-Oxley, HIPAA, or whichever regulations apply to them. They must also ensure they stay up-to-date as the regulations change over time.
Government: Government regulations need to be complied with, and, again, companies operating in two or more geographical regions will need to comply with the regulations for each region. It is important that someone is keeping informed of new regulations and making sure they are applied. Failure to do so can result in huge financial penalties.
Flexibility: Established companies can begin to find themselves ossifying and unable to speedily take advantage of new opportunities. It is important for companies to remain flexible so that they can take advantage of new hardware and software that becomes available and business opportunities as they present themselves. SOA may seem like an IT initiative, but should really be a way for the whole company to re-invent the way it does business - a business modernization project. This is just one example of an IT initiative that should really be a whole-company initiative driven from the boardroom.
Q4: Provide your predictions of future IT trends and their implications/opportunities?
Trend 1: (Dare I say) Service-Oriented Architecture (SOA)
SOA has been an exciting "new" idea for about three years now, but we are finally getting to the point where ordinary companies are in a position to actually implement it in a way that is beneficial to the business rather than just an exercise in IT. Unless you've been living on the planet Zog for the past three years, you know that SOA works by exposing bits of code (but thought of as business services) from existing applications and using them in new Web-based applications. CICS (particularly with Version 3.1 and the recently-announced V3.2) is very good at this. The important thing is that when considering what should be exposed, it should be considered in business terms rather than programming terms. SOA has been said to stand for Same Old Architecture because the idea of services has been with us before. Prior to Web services, we had CORBA services and Tuxedo services. As well as the mainframe, the exposed code could reside on just about any platform, which can make the management of an SOA environment very complex, and identify where things are going wrong (or running slowly) can be a major challenge. Having said that, companies can save a lot of money by re-using code that already exists and works perfectly than writing and testing new code. SOA has been a bit like client/server was - everyone was talking about it, but hardly anyone was doing it. We will now see many major organizations implementing and benefiting from the use of SOA.
Trend 2: Virtualization
My mainframe background meant that I expect processors to be running at 70% capacity or higher. I was amazed recently to hear that x86 servers tended to run at about 20% capacity. Also, companies tend to know where their mainframes are, which, apparently, isn't always the case with x86 servers! Virtualization (and remember IBM has been selling VM for over forty years so it knows a lot about this topic) allows users to run a little bit of hypervisor code on a chip, which then allows it to run more than one operating system. Looking at these x86 boxes, you could safely run three or four of your existing servers on a single virtualized server. The financial savings in hardware costs are huge. Software is available from VMware, Xen, and Microsoft, and others. IBM's virtualization software allows mainframes, mid-range boxes (System i and System p), etc to be linked into a single manageable unit. It becomes possible to allocate virtual devices to operating systems as required using software. The savings in hardware could be enormous for a company, the benefits in management and control are worth considering too. It is also possible to run development environments using virtual components to simulate the real environment, before moving an application across to the production environment. This can result in large savings.
Trend 3: Security
Almost any year could have security as a major consideration for an organization. Problems like viruses we have come to accept as a fact of life. And, as such, most companies have ways to deal with the problem in place (anti-virus software, firewalls, etc). Many recent reports in the press now deal with wifi problems. Wardriving is where a neighbourhood is searched from a car for wifi hotspots, and unsecured ones are noted for later use. With many people working from home and being able to get into their workplace system, it becomes even more important that the weak link in a company's security isn't a programmer's home wifi set-up.
Companies also need to ensure "pharming" is not taking place. Pharming involves legitimate Web site traffic being hijacked to a different and bogus Web site. User information could be obtained and then used to obtain money or services from the company without the end user being aware of what's happening.
By the time you read this there will be a new security threat. Companies need to be vigilant and ensure they are secure at every level.
Trend 4: Open Source
There was a time when you had to buy software from well-known companies or else you could find that your supplier had been taken over or gone out of business by the time you needed to get back to them. There was a time (back in the 1980s) when CA seemed to buy any small to medium-sized competitor. Even more recently I have found that the number of companies supplying CICS-related products seems to get smaller each year as they are swallowed up by the bigger players. PC software is similar with companies coming and going. It seems safest to use Windows as the operating system and Office as the office application. However, that may not be true anymore and seems likely to be less true in the future.
Open Source software has always had a bit of a mixed reputation. It is typically created by enthusiasts and developed by other enthusiasts. It tends to do exactly what it says it will, and it has been run on a variety of hardware during its development. The problem for a potential user is what do you do if it doesn't work - who do you take it back to? The second problem is what do you do if you can't get it to work? Who do you call for support? These problems are largely disappearing as some major companies are making their products Open Source. IBM's Eclipse is now an Open Source development tool. Sun's StarOffice is now available as a free alternative to Microsoft's product. And, of course, there's Linux, which is an Open Source operating system and an alternative to Windows. Should a company risk using Open Source software? If they look at how much they are paying in licences to Microsoft they could easily produce a business case for migrating to Linux. The real crux comes with specialized software. Is there an Open Source version that would work so easily? Assuming there is or assuming the software actually runs on Linux (in a Windows partition, perhaps, or using the results of the WINE project - www.winehq.org/), it is quite likely that the coming year will see a large number of companies increasing their use of Open Source software. And, in addition, if you change to Linux, you can still used Web-based office suites like Writely (for word processing), etc. Your users could go home and sit at a Windows PC, they would then continue to use the same Web-based software. The final reason for using Linux on your PC is because you could also be using Linux on your mainframe. It would make everything seem more cohesive.
Trend 5: "Green" data centres
Global warming, recycling, carbon footprint - everyone is getting concerned about environmental issues. As a company, it could be a good selling point to customers (let alone the actual good you would be doing) to show that you were taking steps to be more environmentally friendly. A "green" data centre may not be a reality, but a "greener" one is. For example, using virtualization reduces the number of servers a company needs and reduces the amount of heat produced. Virtualizing your servers makes you greener. Other ways include making cabling more efficient so less heat is lost - so use newer components from your manufacturer, who will also be trying to appear "greener". Air conditioning can be improved so that the hot and cold air supplies are kept separate - which makes the process more efficient and greener. Stopping people using cars to travel to and from work can be a greening process. The easiest way is to let people work from home wherever possible. Pretty soon we will be seeing lists giving the "Top Fifty Ways to Make Your Data Centre Green". It is important.
Trend 6: High Availability
No large company and not many middle-sized ones can afford to not be able to provide access for anything other than very short periods of time. Because of this, many companies are looking at High Availability as an option. What this means in real terms is that data from one server is backed up regularly and frequently onto a second server - and data from the second server is most probably backed up on the first. Other changes need to be in place so that, should there be an unplanned outage, users are automatically swapped to the new server. The change could even be invisible to them. It is important that things like IP addresses swap across too and the changeover is seamless. We've all been irritated phoning companies to be told that they can't do anything because the system is down - in future that should not happen.
Q5: Make your predictions for the future (no boundaries or topic limits here…)
- Governments all over the world will attach RFID devices to prisoners to identify where they are (in case they escape). Hospitals will attach RFID devices to all patients, starting with new-born babies. Surgeons will scan a patient to see what operation to perform. Parents will want RFID devices on children so they can see where they are (in case of accidents, kidnappings, stroppiness!). Migrants all over the world will have a RFID tag as soon as they enter a country. Governments will attach RFID devices to all vehicles so they can charge for miles travelled (or, if on a bicycle, calculate some offset value).
- City centres everywhere will enforce congestion charges.
- Flying will become a luxury available to the rich and politicians only. New green taxes will be applied to flying that will make air travel prohibitively expensive.
- On Demand computing will not be the success story people thought. Users will virtualize rather than have unused capacity sitting on their machines.
- Hand-held devices will replace laptop computers as the device of choice for everyday computing. Mainframes will continue to dominate in large businesses.
- IPv6 will become a de facto standard after years and years of waiting.
- Networking implants will be commonplace - touch your ear to receive an in-coming call or announcement.
- Sentient and semi-sentient devices will be everywhere.
- Rising sea levels will reduce the amount of fresh water available and drought wars will be fought. Desalination will become an important and economic technology.
- Micro-generation of electricity using wind, solar panels, geothermal energy, etc, will become accepted and commonplace.
- Wifi will be everywhere and free.
- A personal phone will work as a home phone, and IP phone, or a mobile (cell) phone, depending on where you are when you make or receive a call.
- Mobile (cell) phones will be used to pay bills rather than credit cards because of the amount of fraud associated with credit cards.
Q6: Which are your top recommended resources?
- Wikipedia (http://en.wikipedia.org/wiki/Main_Page). This often-maligned site has excellent up-to-date information on so many topics, and has good links to other sites.
- Google (http://www.google.com). Still the best search engine out there.
- Babel fish (http://babelfish.altavista.com/). Quite often I receive text still in the first language (ie not English) of the author. Babel fish is usually able to translate it (or if not, certainly give me the gist of what's trying to be said). A very useful tool.
- The National Gallery (http://www.nationalgallery.org.uk/). Nothing to do with computers or computing, London's National Gallery Web site is an excellent site and not only shows you the brilliant paintings in its collection it tells you about them.
- The Arcati Mainframe Yearbook 2007 (http://www.arcati.com/newyearbook07). The Yearbook contains a mainframe user survey, vendor directory, a media guide for IBM mainframers, a glossary of terminology, articles, and technical information. Everything a mainframer needs in one place.
Q7: Provide commentary on three topics of your choosing.
Topic 1: Project ECLipz
IBM's project ECLipz is an interesting piece of viral marketing. By that I mean, although IBM has denied the existence of Project ECLipz, information about it has passed from person to person until almost everyone has heard of it.
The letters "ipz" at the end of the name are meant to refer to IBM's System i, System p, and System z servers. The thinking behind the project is that it costs money to design and build a computer - and it costs three times the amount of money to design and build three separate (i, p, and z) computer families. Therefore, why not find a way of designing and building everything only once and cutting your total costs by two-thirds. Now, this works quite well for chassis and memory, but not quite so well for the actual processors. Although, having said that, the System i and System p processors can run on Power 5 chips. Mainframes are quite different and can have their own specialized zIIP and zAAP processors etc.
If IBM goes for a 75% sharing solution, they have cut their costs hugely and Project ECLipz has been a success financially for them (although, of course, in the tradition of the cold war they deny its existence). However, some people are suggesting that the new Power 6 chip, when it finally appears, may be able to run mainframe operating systems - making the ambitious project a complete success. We will know by the end of this year (2007) whether they achieved their ultimate objective.
Topic 2: On Demand virtualization
On Demand computing describes a method of making computing resources available to users as they are needed. Basically what happens is this, users have lots of computer capacity available, but pay only for the capacity they use. That way they are not paying to support peak capacity during quieter periods. On Demand computing has been tipped by many to be the next big trend in computing. After all, it makes sense for a company to have all the capacity it will ever need, but only pay for the capacity it is actually using.
However, I am suggesting that computer trends won't move in that direction. I'm suggesting that with the growth in the use of virtualization, users will be able to get more out of the hardware they already have on their site and won't need to have a large amount of just-in-case capacity. Virtualization will reduce the amount of hardware that needs to be switched on, and so will reduce electricity bills. On Demand also requires that data centres have the hardware delivered and installed. Not having it delivered will save petrol and will be a component of a company's "green" strategy. Companies will also need less air conditioning (and again less electricity) because they won't have to cool these extra boxes. The big thing in favour of On Demand computing compared to virtualization is that it is an easier strategy. The extra boxes are used when needed and money is paid out at the end of the month (or whatever the charging period is). With virtualization you need to buy the software and have someone who knows what they're doing to install it. You then need expertise to set up and run the virtual machines. I still think the advantage of virtualization is so great that it's growth will see a reduced take-up of On Demand computing.
Topic 3: AJAX
Google Earth (earth.google.com) uses AJAX, and Flickr (www.flickr.com) the photo sharing site uses it too.
Companies like HostBridge are looking at CICS and AJAX. They suggest that AJAX enables true two-tier access to CICS (rather than three-tier). They say that AJAX allows the browser to contain the application logic and make calls directly to CICS.
William Data Systems has a product called Ferret that is able to monitor Enterprise Extender. What makes it so interesting is that a user from a browser can access so many monitor "pages" so quickly - and it is possible because the product uses AJAX.
I predict that more monitoring products will start to use AJAX as a way of very quickly presenting data to users.
Closing Comment: Trevor, we thank you for sharing your time with us and we will follow your work in the wider international community with interest.
A: Always a pleasure - thank you.