So, What is a Service?

In this day and age with Service Oriented Architectures being all over the place (apparently) it doesn’t seem like the sort of question that I should really be asking. However, I do ask the question because there are, as always, the Purists and the Pragmatists.

The purists would describe a service as being some piece of code that accepts some form of input, performs a function on the data it is given and then produces the results for you. Typically the input and output are in an XML format because it would be “bad form” to expose a service that wasn’t using an interoperability standard to communicate with other pieces of the puzzle.

The Pragmatist would open up the whole debate around Service orientation by questioning the limited view of a “Service” that the purists take. Refuse collection is a service that we all use as are utilities such as water, gas and electricity. Getting a package delivered from one side of the planet to another is also a service that anyone can use from large corporations to the likes of you and me.

The difficulty in getting SOA to “bite” is that the business already sees itself as providing services to its customers as well as its employees, but doesn’t fully appreciate the fact that it is also providing a service to itself internally. That means that businesses don’t fully concentrate on the efficiency of their internal processes or they believe that purchasing a monolithic piece of software will solve their problems. How many organisations have bought a huge HR or Finance system and in the interests of economy have determined that they will shoehorn their organisation into working the way that the software does? After all, we have the term COTS (Commercial off-the-shelf) software – by implication the business will install the software and expect to adapt their processes to the way the software operates.

One of the worries that I have is that we have moved from bespoke, tailored solutions which fit the business like a glove (or at least the glove as worn 12 months ago when the analysts were looking at the business) to a “one size fits all” (OSFA) model. The OSFA model makes a huge assumption that the software will be built to take into account current “best practice” and that may well be true, but “best practice” by whom and for whom? Some time back – in my history – I worked for a database vendor who created applications to support functions such as HR, Finance etc. The problem with the solutions that they had created was that they were ideal for a Database Vendor and as such didn’t really fit healthcare businesses or government establishments particularly well. As a result, after spending inordinate amounts of cash on systems, they needed to be customised to fit the way that the business really worked.

Clearly the OSFA model is not the best for the end customer, but of course it is a really good model for the software vendor.

If you can now take the business and look at it in a component fashion and determine what the smaller building bricks are, then you should be able to start making good use of the SOA paradigm. Take out the monolithic applications and create “Services” which may be performed by people, resources or automated technology. If you can now provide smaller components of technology that will support the people and resources that are trying to provide those services then the amount of bespoke work that needs to be done is reduced. A “Service” becomes something that a capability unit within an organisation provides for the rest of the organisation or the customers of that business. Now when the organisation adapts to external and internal influences the effort in adapting is down to changing smaller pieces of the machine.

Cut and Paste – the key to IT service evolution

One of the most important innovations within the computer world was not the PC in and of itself. Even the ability delivered by graphical user interfaces such as Windows to have multiple applications open simultaneously and to be able to switch between them at will without having to close them down was not the key. Cut and Paste – that was the key to the PC revolution.

Cut and Paste gave every user the ability to move data from one application to another even though there may have been a need to re-arrange the data to get the results that you were looking for. Cut and Paste gave users the ability to look at text based mainframe systems and copy data from that screen into spreadsheets allowing them to create reports that used the data in ways that only they understood.

Data started appearing in presentations that had been massaged from legacy systems, online systems and printed reports and the business user began to get more involved with the analysis of the business rather than simply reporting results. The key factor here in terms of functionality was that the end user really didn’t need to understand the format of the data (was it Binary, Text, HTML etc) because the Paste function would provide a default interpretation of the data. In essence this was the first attempt at interoperability that was commonly available.

In today’s world the business user is still making full use of cut and paste as a means of aggregating data sources and engineering their own data interoperability. The actual work is now being done by personal assistants and secretaries or the computer savvy end user.

The current thinking in the IT world is around “Service Oriented Architectures”, a paradigm that claims to put Business Processes at the forefront of IT thinking. It can be further defined as the mechanism for understanding the business by breaking it down into smaller processes which can themselves be broken down into a set of simple steps which may or may not require an element of programming or computer interaction.

Each of the steps can then be described as a “service” (hence the Service Oriented Architecture – SOA) which has some clearly defined inputs and some form of simple activity which generates a clearly defined output. In essence a “service” is the Cash Machine of the IT industry. You key in your PIN number and how much money you want and after some clever work behind the scenes your cash and receipt are delivered from the machine – Input, Process, Output.

The problem with SOA is that it requires IT staff to engineer the weaving together of these “services” to form a coherent process because of the problems of trying to pass data from one service to another. In essence the problem is now back down to the ability of services to interoperate or specifically, the end users to be able to copy the output from one service into the input for the next service.

For SOA to become the next revolution in IT there is going to need to be the equivalent of Cut and Paste – an easy to understand and use interface for the end user that allows them to copy the output from one service and paste it as the input to the next. Providing the end user with this type of capability will enable SOA to be something that is handed across to the business to design rather than remaining the domain of IT specialists.

Why has SOA Failed to Deliver?

I was having a conversation with someone just the other day about SOA (Service Oriented Architectures) and I think we had an interesting discussion about why SOA has failed to deliver on its promises.

If you look at SOA in its purest form, it would seem that Service Orientation has been around for some time. Let’s face it, the concept is not a new one and I have been around the block enough times to know that I have seen it all before. I like to make the comment that SOA is, in many respects, Procedures and Subroutines in “old money”.  In its simplest form SOA is a black box which takes some XML input and spits out some XML output – so it is a procedure/function with parameters. Given that rather more cynical approach to SOA then it isn’t really something new.

If it really isn’t that new then why is it that it fails to deliver in the current environment? Surely the concept is something that we should all be aspiring to, especially given that the decomposition of everything in sight down to bite-sized chunks seems to offer significant benefits to the IT practitioner or designer.

The conclusions that my friend and I came to is that SOA as a concept seems to be fundamentally sound, it is the fact that it is driven by the IT function within an organisation that really kills its success. In essence I see that there are two warring factions within any standard business – and they are the IT function and the Business itself.

The real key is that Businesses have seen themselves as “capabilities” for some time. Organisationally they are subdivided into Finance, HR, Sales and Marketing etc. Each of the capabilities offer a service to the rest of the organisation and enable the business to function successfully. Each of the capabilities has plenty of scope to do what they want within the bounds of their piece of the organisation and as a result they take ownership of their function. They employ specialists in their field to do and manage the services that they offer to the rest of the business.

Why, then, is the concept of SOA so foreign to the business? I think it comes down to the fact that the IT department are the people arguing for it! There is an inherent distrust of IT departments and a long history of promising and not delivering. Then of course there is also the usual “not invented here” argument especially given that these capability units are now trying to protect their fiefdoms.

The real key then is for the IT to ensure that they have business facing analysts working as part of the capability units and seen as part of the business unit. These “Architects” will be able to become a significant part of the capability unit and thus act as the bridge between the two parties. The type of person required though will be someone who needs to have strong negotiating skills to enable them to broker the appropriate deals with the business. They will need to fully understand the business and should be able to see it in the context of the whole enterprise.

In short … I think that SOA needs to be Business led and Business driven. Having said that, the IT department can make sure that the agenda is driven subtly forward by “infiltrating” the Business Capabilities. It doesn’t sound like a major “revelation” but in the same way that systems in the past have failed because of a lack of communication with the “user”, the same is true of SOA.

Night Flying

Right … I’m on the path to get my Night Rating and so this week I set off having already completed 1h 20m of night flying to do some more. The next step in the process was to do some circuits with various emergency situations imposed on me. I did seven circuits in all and the challenges were interesting to say the least.

1. Standard circuit with nothing unusual.
2. Standard circuit again with nothing unusual
3. This time the PAPI’s were turned off – these are the four lights to the left of the runway (at Southend) that help you with your glidepath. If there are two white lights and two red lights then you are on the right glidepath. More red than white, you are too low, more white than red and you are too high. The runway edge lights can be used as indicators as they merge together (when seen on approach) towards the far end. If they are all merged then you are too low, if they are all separated then you are too high. So … no PAPI’s and eveything went swimmingly.
4. Next time round, no PAPI’s and no centre line and bar extending before the runway. Again – we had a neat landing and everything went well.
5. This time around we had the runway edge lights dimmed right down which really didn’t seem to make that much difference apart from being a little harder to see the runway while we were in the circuit.
6. The final challenge was when the instructor reached across on the turn onto base leg and turned off all the instrument lighting in the cockpit. I couldn’t see the altimeter, Airspeed indicator or any of the other instruments. Out came a torch which I placed on my lap and was pleased to see that I had managed to get the aircraft into the right attitude and airspeed purely from the sound and look of things. Again – no mishaps.
7. inally we went round again and did a circuit under normal circumstances.

A quick backtrack on our full-stop landing and we were off for a debrief and a swift pint.

I thoroughly enjoyed the challenge!

The Flying God

I’m not sure how this is going to work for me or anyone that cares to read this but I wanted to start to log some of the things that are important to me in this type of forum. I’m sure this is the way that most people start out -what do I say, how do I write it … will anyone even care?!

In the last twelve months I managed to get my Private Pilots License in the UK – flying from London Southend Airport. I’m now embarking on the next step in the journey, and that is to get my Night Flying rating. I currently fly a Cessna 172 (a four seater aircraft) and although it is an old aircraft it has served me very well.

The last flight that I took was my first 1h20m flying at night. We flew south over Chattham in Kent and then followed the M2 towards the QE2 bridge. I was amazed at the sights and I absolutely loved it – it was a whole new experience from being up there in the daylight. From the QE2 bridge we went and flew over Basildon, Chelmsford and then out to Colchester and returned to Southend. It was a relatively un-remarkable flight, in the sense that all of the flying was pretty standard, it was the whole experience that was incredible. I’m trying to think of the right words to describe how I felt – there was an excitement that was new – was that just the new experience of being up there at night … Hmm … I’ll have to mull that one over.

On the technology front, I have been playing with Google Documents and the Calendar etc today. I am a big fan of Open Office just because I find that it is properly integrated. Did you know that you could upload Open Office files to the Google Spreadsheet and Google Document sites? That worked pretty flawlessly – just struggling with the printing piece, but then in my day it was always the printing piece that caused the most trouble!

I’m going to leave it at that for today. I need to go and make my favourite cocktail – The Caipirinha (pronounced Kai-pi-ri-nya). Basically it is Limes and Brazilian Rum with a little bit of sugar in the form of Maple Syrup. Yum!