.comment-link {margin-left:.6em;}

Sam Lowe's blog on Enterprise IT

Wednesday, January 17, 2007

Architect, Functional and Technical: IT's Good, Bad and Ugly?

Despite most Enterprise IT people knowing the downsides of silo-thinking in a business, many of us are aware that IT can often act as the biggest silo of all within the business of which it is part.

The ultimate objective for many is to run IT like an independent business (commercially-driven and efficient demand, supply and resource management). But there is a trick to doing so without it starting to think like a separate organisation (obsessed with its own bureaucratic processes, pet projects and irrelevant agendas). Getting the good without the bad is a challenge that has led many to embed IT back into the business units as an alternative solution.

But there are also cliques within IT itself which tend to exist whether it is centralised or embedded. The differences between the 'management' people (focused on delivering on time and to budget), the service people (focused on keeping the kit running and the lights on), and the 'content' people (focused on the solutions and the design) are obvious. But the latter in particular itself has silos. Those of functional, technical and architect.

The functional-technical divide goes way back. Even now, organisations still tend to have separate functional and technical teams. It still seems rare for anyone to really bridge the gap. One of the great characteristics of packaged apps was that they gave frameworks through which the technical and functional teams could interact (often agnostically) in more predictable, industrialised ways. This brought them together more efficiently in terms of delivery, but did little (if anything) to bring them together in mindset.

Architect-types have sometimes thought that they could cover the whole spectrum, and also aimed to bridge the portfolio-level (the enterprise view) to the application-level (the project view). But very often this just has created a third silo - the so-called ivory tower. Additionally, with many architects having come from infrastructure-planning or bespoke-development backgrounds, in the many industries dominated by packaged apps they have had additional disadvantages in building the credibility and buy-in they've needed.

However, today's increasing focus on SOA and in general architecture above the level of the project, driving cross-project synergies and initiatives requires that the three viewpoints work together. Take composite applications for example. How can (what are by definition) composites of data, function and technology assets, mixing packaged and bespoke, be put together scalably without united functional and technical views? And how can a portfolio of them be managed without an architecture view across them that balances the delivery-proposition to each project, with the customer-proposition to each part of the business, and the ongoing ownership-proposition to the enterprise as a whole?

None of these issues are (in principle) different from those of our recent pasts, but this is the first time that the very value proposition of the approaches themselves have been contingent on these viewpoints being able to work together. And making it stick.

Technorati Tags:

Sunday, January 07, 2007

The Inmates Taking over the Asylum? Traditional Data Management Meets Web 2.0

What does the way the public internet is evolving tell us about how Enterprise Data Management might change?

After having presented on these questions recently at DAMA Europe, I've been contacted several times about running an independent follow-up workshop in the London area in the early part of 2007 to work through some of these issues. Anyone who'd be interested in such a workshop, please drop me a line through a comment, link or the email address on my profile page.

Enterprise IT has always looked to improve efficiency through industrialisation and automation. This has often meant standardisation, centralisation, and consolidation were/are used as methods of enabling this. Enterprise Data Management has been no different and has generally followed this path. Whether you're thinking about (1) relational databases and data modelling, whether you're thinking about (2) data warehouses, the CIF and business intelligence, or whether you're thinking about (3) metadata management, business activity monitoring and corporate performance management ... all generations have held standardisation, centralisation and consolidation either at their core, or at least relied on it to a significant degree.

But outside of the enterprise, the internet of enthusiasts and consumers has had rather different dynamics. Rather than automation and industrialisation, its key objectives have tended to be more like access and collaboration. This difference has enabled interesting emergent innovations in the internet that enterprise IT has been able to absorb back into its own practices.

This applied to the original world-wide web and the eCommerce wave where the mass access to information at almost zero cost enabled new channels. New channels for organisations to interact with their customers and suppliers, new ways for systems to interact with their users, and new ways for employees to find and get at information hidden in their organisations' islands of information.

It's interesting to think about what the current wave of evolution in the internet (generally referred to as Web 2.0) might mean. One of the characteristics about what is known as Web 2.0 is that rather than the web primarily acting as a new channel providing greater access and cheaper communication, the emphasis is on the web being the platform itself, providing a basis for new levels of collaboration and cheaper participation.

For the Enterprise Data Management world there is an additional focus in this. As with the original web generation, although the channels available and information access grew, the data was generally still 'owned' in the same place it always had been. However now, the Web 2.0 evangelists and analysts talk about the user controlling their own data - which should grab the attention of all professionals involved in enterprise data, as it's a big mindset change from where things have been in the enterprise.

There are many interesting patterns in Web 2.0 that warrant analysis for clues as to how enterprise data may evolve. The Long Tail for example is very relevant, as enterprise data's current techniques and approaches are generally very expensive in terms of labour, time, and politics and so can only be applied economically (if at all) to the most common pieces of shared data, and not the huge amount of other data. The dynamic bottom-up network effects and the emergence that follows fosters behaviours and insight not normally available with the top-down governance and compliance methods of traditional enterprise data. And the user-provided tagging is very different to the rigid and controlled hierarchical taxonomies, structures and syntax that typically have been the basis of enterprise data approaches. And so on ... (I'll follow up on some of these separately).

So Web 2.0 will undoubtedly introduce new technologies like RSS, Wikis, REST, Blogs, AJAX etc into enterprise usage. But perhaps more interesting than the new technologies is the potential influences it will have on Enterprise Systems design and the way Enterprise Data Management will adapt.

For example:-

Web 2.0 EDM InfluenceChallenged EDM Characteristic
Intellectual CommonsNot just Elite Experts
EmergenceNot just Upfront Grand Design
Dynamic MetadataAs well as Fixed Syntax
Tagging & SemanticsInstead of Just Hierarchies
ResourcesBeyond Just Data
LinksAs well as Replication
Others’ Resources as Your PlatformAs well as Your Own


Many will recognise that there has been quite a lot of discussion about what Web 2.0 will mean to Enterprise IT generally (Andrew McAfee's Enterprise 2.0 threads are a fine example) but I havn't yet come across discussion on what it specifically may mean to Enterprise Data, which would seem to be one of the prime areas affected.

Technorati Tags:

Saturday, January 06, 2007

Architect Insight 2007 Event

Matt Deacon of Microsoft UK and IASA UK has just blogged about how he's tried to construct Microsoft's Architect Insight Conference at Celtic Manor in March this year so that it is as interactive and inclusive as possible, with discussions and working groups that are deliberately intended to ask questions and work through issues rather than presenting predetermined solutions. I'm looking forward to this dynamic, it should produce some interesting results, it might reduce the level of death-by-powerpoint, and maybe even keep some of the 'spin' in check.

What's particularly unusual about this event is the breadth of the agenda, with work-streams covering a broad set of areas such as 'enterprisey' strategy issues like the one I'm speaking on, through web, security, and infrastructure, as well as the software design issues (that you'd perhaps expect from an MS -organised event). It makes for an interesting mix of perspectives.

Technorati Tags: