At this very moment I am stuffed in 22F on a completely full flight returning from New York and it gave me a chance to look at my journal, and review the notes from a couple dozen or so conversations I have had this week and last. While none of my conversations were specific to EA, there were a number of very consistent and complementary themes which will, I suspect, impact the EA of complex organizations over the next several years.
First, the conversations.
Yesterday, I spent several hours with Hardev Dhindsa who has a small firm, Exact Solutions, whose iWatch technology is installed on roughly 2,500 performance critical servers in 7 of the 10 largest investment banks (e.g. Lehman, Goldman, JPMorgan, Barclays) and no where else – let me repeat – “no where else”.
I was introduced to Hardev by a mutual friend who thought I could give him some insight into the potential market for his product and help with strategic planning. As that conversation progressed, Hardev and I began talking more and more about the evolution of IT (in the broadest sense) over the next few years.
(At some point I will write more about the evolution of ‘systems management architectures’ but for the moment let me say that Hardev has built a query-level performance monitoring tool that is lightweight, non-invasive and used by all of these banks to ensure very high (subsecond) service levels for SQL queries. The simple fact is that tools from CA, Tivoli, HP, BMC, et. al. provide effective OS monitoring, database monitoring and network monitoring but they don’t provide insight into the performance of SQL queries. Nor do these tools help ‘support staff’ identify which queries create problems. Hardev’s approach is as brilliant as it is simple. And DBAs faced with unrealistic pressures to ensure performance should embrace iWatch to simply prove – “Hey, it’s not a database problem!” But that’s another blog.)
On Monday, I spent a couple of hours with Harold Heath who is Enterprise Technical Architect at Citigroup and someone I have chatted with for awhile. In the course of that conversation, Larry Burgess, another EA at Citi who I have met before, joined us.
Last week, I was at the GoldenGate User Group Conference where I delivered the closing keynote. (In the interest of full disclosure, I have been on GoldenGate’s Board for about four years and I think they are the leader in transactional data management.) I had the chance to speak to a number of GoldenGate’s clients and partners. Attendees included end users such as UBS, JPMorganChase, Visa, BofA, First Data, Shell, Sabre, MGM, ADP, etc. and partners including IBM, HP, Teradata, Ingres, GE Healthcare, Cerner, Amdocs, etc. The conference had one of the best CIO panel discussions I have ever heard. On the panel were Roger Burkhardt who was CTO at NYSE for 6 years (before recently joining Ingres as President & COO), Penny O’Hara who is CIO for BT Health which is leading the re-engineering of healthcare delivery across the UK, and Craig Murphy who just retired as CTO of Sabre Holdings.
Now to my journal.
So, I talked with and listened to a bunch of really bright people. At first my notes appeared to be all over the road but I eventually reduced them to a consistent set of nine, very concise, ‘EA Futures’. After reviewing my first draft I decided that my desire for conciseness might, unfortunately, demonstrate the Law of Unintended Ambiguity (I just made this Law up), so I thought a little ‘translation’ might clarify as needed.
Here we go…the 9 EA Futures to come out of my recent conversations are:
1. ‘Electron-ification’ of all business processes is the goal. Translated as “straight-through processing in as ‘real-time’ as we can make it is how we want everything to work.”
2. Self service everything. Translated as “most users that need to interface with #1 should just do it themselves any time they want; they get better service and we reduce headcount, wahoo!!”
3. Active data warehouses. Translated as “‘electron-ification’ requires ‘self-service’ which means information has to be current not just accessible”; said differently, “we can’t do #1 and #2, unless we do #3”.
4. Pan-process/pan-service. Translated as “we need to take an enterprise view of everything (i.e. #1, #2 and #3”) whether we are talking about business processes or application services”.
5. ‘XML-ification’ of the enterprise. Translated as “this is how #1 thru #4 talk to anything most of time, and talk in a very agile way, but XML has a lot of overhead”.
6. Horizontal scaling is a must. Translate as “#1 thru #5 means way more bandwidth, way more storage and way more processing so you better think in terms of building out and not up”.
7. Cost-effective computing is required. Translated as “relentless focus on driving down the cost of terabytes, gigabits and MIPS in the context of #1 thru #6 and the requirement that you bow toward San Francisco everyday” (note: San Francisco was Gordon Moore’s birthplace).
8. Open systems is a key strategy. Translated as “Linux” or “you can’t get #7 done without it, and if you can’t get #7 done forget about #6, and if you can’t get #6 done you can’t afford what’s really required for #1 thru #5.”
9. Zero downtime. Translated as “planned outages are just as bad as unplanned outages so you better correctly architect and enable #6 thru #8 so you can deliver #1 thru #5”.
Yeah, it was a stimulating two weeks. And this blog barely scratches the surface.
Hold the Date! – George Paras Keynote at Denver IASA ITARC on May 6, 2010
February 28, 2010In a keynote presentation at the upcoming IASA ITARC Denver Event I will be discussing the leadership challenges that many current and aspiring IT leaders face in their organizations, and how they can overcome those challenges by applying selected enterprise architecture concepts to their decision-making framework. Not surprisingly, I advocate moving EA “upstream” within the management structure and expanding the context to make it more relevant to the business. Doing so isn’t easy. It requires a combination of skills in leadership, process integration within the Office of the CIO, and appropriate content in business and information architecture. In this session attendees will learn how to get started on this journey.
If you are in Denver and plan to attend, please let us know. We’ll be available to discuss ideas with any organization interested in expanding the scope and reach of their EA practice into business-oriented EA.