Breaking the Code

Controlling the Weather

Hello Muddah, Hello Fadduh, Here I am at Camp Granada ...

Well actually, not quite, I'm currently on vacation in Fort Lauderdale, Florida with my family.  (Well strictly speaking my family is on vacation there and I've sneaked up to NY for a couple of days to catch Deutsche Bank's FinTech 2008.)

Anyway the point is that in Florida the weather is a major topic of conversation and a more or less permanent feature of local news bulletins particularly at this time of the year which marks the start of the hurricane season.  In fact, there is already extensive coverage of the first hurricane of the season - Bertha - although this looks like being nowhere near as severe as its namesake from 1996.

Of course the joke is that the English love to talk about the weather (if you don't believe me check out online English primers such as the English Club!) which in the UK is pretty benign whereas in this part of the world - Florida, the Bahamas, the Caribbean and so on - knowledge of what the weather is doing or how it may develop can mean the difference between life or death.

In the short time we've been been in Florida there have already been some spectacular storms which got me thinking about Cloud Computing.  Sad but true.

The background to all this is that I have been following the debate led by William Fellows at 451 on its merits and relevance to the enterprise which he addresses in his latest report Partly Cloudy - Blue-Sky Thinking About Cloud Computing.  (Unfortunately I will still be on holiday but you can catch Will at the upcoming Cloud Camp London organized by the guys at CohesiveFT.)

So where have I got to in my own thinking?  And here I have to confess that I was unusually skeptical when Cloud Computing first surfaced having lived through Grid and then Utility Computing yadah yadah.  My concern is that the half life of a marketing concept is now so short that you barely have time to flesh out one and start to ask the really interesting - and tough - questions before it becomes passé to even talk about it!

Well, looking beyond the marketing hype and continuing in the same vane (sorry that was awful even by my standards!), my view is that without adequate climate control and ultimately the ability to control the weather resistance to Cloud Computing will dramatically slow its adoption by business, for example, the financial services industry which would be a huge lost opportunity.

In short the key to realizing the benefits of Cloud Computing in the enterprise is to implement robust service governance to ensure that it is utilized in a tractable, compliant way i.e. by enforcing the appropriate operational risk management policies.

On that topic more anon and hopefully a demo before too long and definitely by HTS 2008 in Vegas.  For now it's time to get back to my vacation.

Intelligent Design My Ar*e!

So the Gartner IT Infrastructure, Operations & Management Summit is winding down.  It's been a hectic few days.  You know that end of conference feeling when everyone is de-mob happy and you really feel for the guy on the graveyard shift (I know I've been there!).  Well this time it is David Williams flying the flag for IT Process Automation so I'd better be there!!

Talking of automation one of the recurrent themes of the conference has been that blindly automating an existing procedure isn't much use - you've got to understand the process you are trying to implement holistically.

I'll give you a great example -

I travel a lot and last Friday I found myself at Heathrow en route to Edinburgh for a stopover to see my family, remind them I exist etc.  Anyway I pitched up at T3 and because I was changing airlines I had to collect my baggage then make my way to T5 rather than stay air side.  So I got the train and when I came out of the underground station I went up the escalator and ended up in arrivals by mistake.  No problem I thought and I followed the signs until I found the lift (elevator) I should have taken from the station to departures.

Instead of a pair of up/down call buttons there was a  sign saying "Lifts stop Automatically" which had been crudely sellotaped to the panel.  Bit tacky I thought but hey its early days and at least they've implemented a smart lift.  So I waited.  The first lift arrived and the light above the doors indicated it was going down so I waited for the next one.  This happened a couple more times so I took a closer interest in the lifts.  Since these were in a funky glass & steel housing I could see lifts going up but for some reason not stopping automatically on my floor.

Eventually I figured that for some reason there was a bug in their system so I got a lift going down and sure enough it then went up automatically to the departures level bypassing the arrivals floor altogether.

As I got out I bumped into a BAA staff member and said "I think there is a problem with your lift.  It doesn't appear to be stopping on the intermediate level going up".

She said "No. There is no problem.  It is designed to take people from arrivals down to the underground station and from the underground station to departures.  Look there is a sign in the lift that explains this." 

Sure enough there was another crudely sellotaped sign on the panel in the lift with a diagram illustrating that the lift went from floor -1 to +1 then from floor +1 to 0 and on to -1.

When I had the temerity to point out that being stuck on the 0 floor I would have to have got into a lift going the wrong way to see this sign she shrugged her shoulders and said "That's BAA for you!" and walked off.

In the words of that redoubtable character Jim Royle from the Royle Family, "Intelligent Design My Arse"!!

Drinking your own Champagne

I am writing this at the Gartner IT Infrastructure, Operations & Management Summit in Florida.

This is the second year Gartner has run this event and for an ISV in the IT Process Automation space it is the Gartner event so we are here in force along with some of our competitors as well as one of our OEM partners Unisys Corporation.

Unisys have licensed our Enigmatec Management System (EMS) as Unisys uOrchestrate™ which is a key component of their Infrastructure Management Suite.

While this is hardly earth shattering news what is worth noting - and the reason for this post - is that Unisys are drinking their own vintage champagne and have deployed our Enigmatec Virtual Orchestrator (EVO) internally.  In the words of Brian Ott (CTO, Systems & Technology, Unisys) they are "now able to deliver 95 percent of requests for virtual machines within ten minutes, completely hands-free.  EVO has transformed our business and dramatically improved the speed at which we deliver solutions to our customers."  (If you happen to be at the Gartner show you can catch Brian Ott here.)

Rather than rely on our ROI calculator Unisys commissioned their own case study which estimated that this deployment will save them $3m over the next three years.  This is modest compared with the savings of some of our large enterprise customers but is nevertheless an important milestone for Unisys themselves.

The bottom line is that, while we have proven solutions now in production around the world managing mission critical business services, as we move forward strategic business partnerships are vital to our growth and, in turn, the internal adoption of our solutions by our partners is key to their success.

In the case of Unisys the speed with which they have embarked on their internal roll out and the immediate payback they are seeing demonstrates that drinking your own champagne is definitely in.

In line with current TTB policy it should be noted that this posting is not intended to encourage you to drink or to increase your alcohol consumption for health reasons.

The Problem with Blogs

It has been a while since my last post.  Well over a year to be exact.

The problem with blogs of course is that if you are not careful a week or two passes between posts then a month or more and before you know it you are in that classic situation where you are too embarrassed to blog!

It's the same with people - you bump into someone you haven't seen for a while and agree to meet up for a drink then leave it so long - having sworn blind that you'll call them the very next day of course - that you can't bring yourself to call.  That's how people drift apart and that's my excuse - feeble though it is - for leaving it this long.

So what has galvanized me into retrieving my account, recovering my password and logging on to Typepad after all this time?  To find out you'll have to read my next blog.

In the meantime I'll leave you with one of my all time favorite poems - Days by Philip Larkin

What are days for?
Days are where we live.
They come, they wake us
Time and time over.
They are to be happy in:
Where can we live but days?

Ah, solving that question
Brings the priest and the doctor
In their long coats
Running over the fields.

Cool poem?  Philip Larkin was also a pretty well known jazz critic in his day.  I think he would have liked Kate Edmonson who I caught at Eddie V's in Austin last Wednesday.  How is that for serendipity?

What is Progress?

Back in February I asked the rhetorical question Where has all the snow gone?  and floated the idea of a new metric Return on Environment (ROE) which I proposed should be used to determine the overall impact of technology innovation.  At the time I argued that ROE should be factored into any overall ROI calculation.

Over the past couple of months I've socialized this idea and got some very interesting feedback from friends and family and now think that ROE should supercede ROI in much the same way as the guys at  Redefining Progress have argued that Genuine Progress Indicator (GPI) should supercede GDP as a more accurate measure of economic and social progress -

GPI starts with the same accounting framework as the GDP, but then makes some crucial distinctions: It adds in the economic contributions of household and volunteer work, but subtracts factors such as crime, pollution, and family breakdown.

ROE needs to take account of the resources consumed both creating and running IT infrastructure; whether these resources are renewable; the cost associated with extracting/acquiring these resources and so on.

In other words the ROE of a complex entity such as a next generation data center (NGDC) will itself be complex and will therefore need to be decomposed into its constituent parts (interestingly in much the same way that the management of such an entity needs to be unpacked) with the overall ROE a function of the constituent ROEs.

Given that ROE typically needs to be computed over the lifetime of a long running entity such as an NGDC we need to find a way of expressing its net present value (NPV) in the same way that we compute the NPV of a cashflow for a complex financial instrument such as bond with embedded optionality, where the latter represents, for example, a specific technology refresh.  In place of interest rates there will be a set of forward rates for resources based on their projected scarcity/impact on the environment and so.  Thus the NPV is itself a measure of the sustainability of the long running entity.

This link between ROE and sustainability is fundamental.  Redefining Progress recognize this in their mission statement which is to work with a broad array of partners to shift the [US] economy and public policy towards sustainability.

Perhaps Redefining Progress are the right organization to take this concept forward?  Their Chair Claude Poncelet was formerly in charge of Environmental Policy at Pacific Gas & Electric who are themselves taking a leadership role with their energy efficiency programs.

Footnote

Out of curiosity I googled "What is Progress" just before posting this article.  I was amazed by what I found.  Ever wondered what John McCarthy does in his spare time?  Take a look - you'll be surprised!

Virtualized Electronic Trading

Virtualization is set to play a key role in the next wave of exchange systems.  This might seem like a commonplace given the impact virtualization is having elsewhere but there are some subtleties to do with the way electronic trading is implemented that make its application in this area non-trivial.  But before we get into this lets look at the history of electronic trading.

The Rise and Rise of Electronic Trading

An Electronic Communication Network (ECN) is a computer system that facilitates trading of financial products outside of stock exchanges. The primary products that are traded on ECNs are stocks and currencies.

ECNs came into existence in 1998 in the US when the SEC authorized their creation.  There is a good entry in Wikipedia on Electronic Communication Networks which goes into this in more detail.  What is interesting is that the introduction of ECNs and other classes of Alternative Trading System (ATS) has accelerated the adoption of electronic trading by the traditional exchanges.  This has resulted in significant M&A activity with ECNs such as Archipelago and Instinet being acquired by the exchanges themselves.

Some analysts have gone as far as to say that at this point ECNs and exchanges are virtually indistinguishable.  A good example is Celent.

The relentless rise in program trading (driven by the Hedge Fund industry in particular) has further increased the demand for automated electronic trading.

The Need to Virtualize Electronic Trading

Against this backdrop, creating an energy efficient computing infrastructure leveraging virtualization, provisioning and orchestration to virtualize electronic trading and deliver a consistently low latency high throughput non-stop quality of service i.e. achieve a specific set of performance service level agreements (SLAs) has become paramount, particularly as the many of the exchanges are seeking to reinvent themselves as service providers hosting third party regional exchanges as well as their own core exchange services.  In Europe a good example is Atos Euronext Market Solutions.

The Challenge

However to create this kind of infrastructure and virtualize an electronic market requires an understanding of the way in which electronic trading works in practice.  In any given market, each instrument (stock, option at a particular strike price, etc) traded electronically has its own order book.  Each order book is maintained by a book manager which is responsible for implementing whatever trading rules are in place (inverting, crossing, matching, sweeping, etc) as well as the top of book (best bid/offer) and depth of book (active bids and offers in price/time priority).  Furthermore orders or adjustments to orders sent by any trader (human or program) must always be handled in the order in which they are submitted by the trader and must never be lost or duplicated.  This is called preserving partial ordering and is a fundamental requirement.

Therefore the moment you start to think about dynamically reallocating resources and migrating order books from one resource to another you have to ensure that you can guarantee partial ordering i.e. the partial ordering SLA cannot be violated under any circumstances.

This is a special case of the general problem of dynamically distributing a mediated service that can be logically partitioned into well-defined segments that applies to not just electronic trading but also to betting, online gaming and so on.

How to meet this challenge while at the same time taking full advantage of the benefits that virtualization and virtualization management brings to the table is something I'll be covering in more depth in my next posting.

What's he building in there?

Fans of Tom Waits will recognize this as the opening line of What's he building? from his 1999 album Mule Variations.  (Lyrics are here if you are interested.)

The same question could be asked of Cisco.  There has been a definite uptick in chatter around Cisco's Service-oriented Network Architecture (SONA) ; what has become of their Topspin acquisition; and their ambitious and disruptive concept of a virtual data center centered on network virtualization.

For those of you lucky enough to subscribe to 451 there is a great Spotlight article by one of their network analysts Network virtualization: are servers converging with switches and routers?

Hold that thought - more on this later.

Where has all the snow gone?

I'm taking a few days off to meet up with some old buddies for our annual First Friday Chapter Skiing Trip.  This is a key event in our social calendar and the only legitimate excuse for not turning up is that you are no longer with us ;-)

Anyway we are in Davos for the second year running and the question du jour is "Where has all the snow gone?"  Even though it is the first week of Feb the lower slopes are almost devoid of snow (apart from runs that are maintained artificially) and it looks more like April here.  Of course this could be due to all the hot air at the WEF last week but I somehow doubt it!

This coupled with the fact that when I was in California last week everyone was reeling from news that
Freeze destroys 70 percent of California orange crop got me thinking about the need for a new metric for technology innovation Return on Environment (ROE).  The point being that this needs to be factored into any overall ROI calculation.

I did a quick Google and to my surprise the only reference I could find to ROE was in a paper entitled LCA Methodology in the International Journal of Life Cycle Assessment (LCA).  The good news is that while this was being used in a specific technical context - to evaluate the impact (positive or negative) of a production process - it has exactly the right definition to be employed more broadly.

The authors also introduce the concept of an Environmental Management System (EMS) which is neat.  So, for example, the value of datacenter management is not just about efficiency defined as improved utilization but is also about improved power and thermal management - hibernating hardware when it is not needed even if this increases the response time.   The latter is not a problem if you factor it into your overall policy.

This all begs the question is California Citrus industry a victim of its own success?  It has become a world economic powerhouse by shipping product around the globe without any thought for the ROE but has this now come back to bite it?  Of course, the equation is far more complex than this but you get the point.

The End of Grid

In a recent Grid Computing Planet article Grid, Virtualization Get Closer Paul Shread highlights research by 451 Group analysts Steve Wallage and William Fellows who in their 2007 Preview - Grid Computing argue that Grid and Virtualization are converging with the former being subsumed by the latter.

I would go further and argue that Grid is a first generation attempt to virtualize IT resources and deliver a dynamic infrastructure which has been eclipsed by the rapid emergence of virtualization technologies that, appropriately orchestrated, finally provide us with the foundations for a Service-oriented Infrastructure (SOI).

Grid is ill equipped to deal with the challenge of how to “dial in” the requirements of individual applications and services and dynamically compose the infrastructure to accommodate these requirements.

Grid is a hammer that sees everything as a nail - it is a well suited to niche application areas like compute but hopelessly inadequate as a general solution where what is patently required is a more sophisticated top down approach.  One where a grid is just a way of organizing underlying virtual resources in a pattern that is well suited to certain applications.

Forget the notion of an Enterprise Grid – this has given way to the far more powerful concept of a SOI underpinned by the "holy trinity" of virtualization, provisioning and orchestration.

Think back to the first industrial revolution:  For Iron read Grid and for Steel read SOI.  And we all know what happened to economies founded on Iron production.  The End of Grid is just a matter of time.

If you don't believe me then check out the Belwether of emerging technologies Credit Suisse. What are they investing in?   Data center scale virtualization of course.

Check out today's BusinessWire announcement Credit Suisse Funds Scalent Systems Expansion.

Virtualization Management is the Key

Happy New Year!

While I've been on vacation a flurry of articles on virtualization have appeared online reinforcing the widely held view that this is the year that server virtualization goes mainstream and moreover that virtualization management is the key to unlocking its ultimate potential.

Here are just a couple of examples.

Writing in InformationWeek David Strom identified server virtualization as one of 5 Disruptive Technologies to Watch in 2007.  I've lifted a quote from this article which is indicative of the way most CIOs are thinking.

"We plan to use virtual server management to reduce our server support efforts, minimize downtime, and reduce the ongoing costs of server replacement, enabling us to support more hardware with existing staff," said Karen Green, the CIO of Brooks Health System.

Most intriguing in some respects was Kirti Devi's call to arms in Enterprise Networks and Servers Virtualize Now!  Kirti is Marketing Manager for Intel's Server Group.

Intel has been at the forefront of virtualization for some time not only supporting the Xen open source project but providing hardware support for virtualization under their Intel VT initiative.

The bottom line for Intel is that virtualization will accelerate the adoption of its new multicore chips.  Virtualization insulates applications from the underlying hardware and enables customers to take advantage of Intel's new Core 2 micro architecture without having to retool these applications.

A neat side effect is that virtualization also enables customers to extend the life of older hardware while introducing newer hardware alongside this allowing for continuous migration/upgrade of infrastructure.

Finally an unashamed plug for OpenVZ.  As Steve Shankland points out in a recent ZDNet story OpenVZ ported to Sun Niagara servers OpenVZ could race up the charts this year as it finds its way into new Redhat and Novell distros.