The budget drought caused by the recession has becalmed the previously buoyant self-build data centre sector, with all but the biggest banks now looking at managed services or co-location as alternatives to cater for burgeoning demand. Neil Ainger reports on the lively discussions at the fifth Data Centre Excellence (DCEV) conference on 1 October at the IoD Hub in the City, where this issue was touched on, alongside green taxes and more efficient IT and design

Just over a year since the fall of Lehman Brothers on 15 September 2008, some of the leading industry figures in the financial services data centre space gathered at the IoD Hub in London to discuss how the banking crisis triggered by that event has affected the type of budget and data centre projects, which facilities managers can now undertake at banks, insurers and trading houses. The theme was 'out with the old, in with the new', reflecting the need to take heed of the changed circumstances in the sector. The boom of the mid-noughties is now over and simply building new data centres or adding more and more servers is no longer an option to meet the growing demand for data processing capacity, necessitated by the rise of algorithmic trading, and the need to store more customer data for cross-selling purposes and to meet regulatory requirements. Many firms deployed virtualisation and data de-duplication to help them get more out of their existing assets, and this trend continues, but ultimately the data centre as a whole needs to be looked at holistically with computer aided design (CAD) modelling for efficient cooling and other technologies, such as dynamic building management systems, helping to drive up utilisation rates.

The aim of the day is to give data centre managers and operatives, CTOs, CIOs and other interested parties, a comprehensive understanding of the causes of the extra demands being placed upon the data centre and how to meet it, while also outlining the future horizon in terms of green taxes, such as the UK Carbon Reduction Commitment (CRC), and coming technologies or regulatory demands. Business continuity issues were also covered under the remit of the changed environment facing everyone post-crunch.

The day was compered by Ashley Davis, a managing director at JP Morgan Chase, responsible for their global data centre strategy, technology sustainability programme and corporate HQ technology design. He started proceedings off at 9am with a short introduction, explaining why this subject is so important by highlighting the fact that JPMC has over 40 data centres around the world, consuming 50MW of energy at a cost of approximately US$40 million per year, which is equivalent to almost 20 per cent of JPMC's electricity bill. The new demands on the data centre, such as sustainability, cloud computing and so forth, were also mentioned.

Straightened times
The first speaker, Stephen Taylor from commercial estate agents, CBRE, illustrated how the recession has hit the data centre sector with a wonderful graph entitled credit crunch impact, showing how demand has changed across the three major markets - self-built facilities, and retail and wholesale co-location premises - with the Libor rate interspersed over the top, and the dates when Lehman Brothers and Bear Stearns collapsed. The correlation between the stuttering markets and the supply of new storage capacity was very interesting to see. "The boom in capacity between 2005-06 was caused by financial institutions deploying new technologies and ran right up until 2008 because of long lead times and the IT drivers were still there," commented Taylor. "Funding did eventually dry up and the first half of this year saw a sharp withdrawal of demand - take-up was 70 per cent less v 2008 for self-builds, which is a startling figure."

After gauging the impact of the recession on capacity, the next topic up for discussion was people, with Geoff Prudence, chair of the Chartered Institution of Building Services Engineers (CIBSE) Facilities Management Group, reminding the 100 attendees of the need for good training and skills amidst the plethora of technology, otherwise the gap between design and operational practice will never narrow. Its importance in maintaining facilities was also stressed.

Chris Edmonds, a managing director, at Morgan Stanley, who heads up the European distributed computing and telecoms infrastructure groups, as well as being global head of enterprise voice services, next showed how efficiency targets and strategic objectives can successfully be met by good project management, when he provided a case study detailing the bank's move towards a virtual desktop infrastructure (VDI). Chris, who used to work with our compere Ashley Davis back in 1985 when they both started out in the industry, pointed out the drivers and benefits of VDI, such as how it can improve security because employees can't plug USBs into PCs anymore, and traders no longer have to wear shorts on the floor because offices aren't so hot without hard drives there. "The key thing though is that VDI gives you flexibility as its extracts your products and services from your physical location - be aware though that it will add lots of demand to your data centre."

The target at Morgan Stanley is to have VDI as standard by Q4 2010 but there are perhaps 60,000 desktops at the bank so this is a challenging figure, admitted Edmonds. So far 700 have made the physical to virtual (P2V) move, with 2,000 sales desktops to follow shortly. A Citrix Xen desktop with Windows 7 has been specified, using VMware Hypervisor and NetApp for storage and replication.

"Getting the finance is not easy," admitted Edmonds, "but you need to focus on running costs, depreciation and so forth [to get sign-off]. I think the entire industry is heading to VDI as well, it's not just us," he added. A lively Q&A session ensued, touching on the role of cloud computing in VDI, support costs, recycling old PCs and other matters, before coffee was taken.

Upon returning Ashley thanked his old colleague, before introducing Owen McKee, country manager, UK & Ireland, at Avocent, who talked about how to optimise your operation, via effective management that allows you to do more with less. He highlighted the usefulness of the Power Usage Effectiveness (PUE) and Data Centre Infrastructure Efficiency (DCiE) metrics, particularly in regard to green IT and driving up utilisation rates. Continuing the theme, Professor Derek Clements-Croome, of the construction engineering department, University of Reading, gave an interesting presentation next on the importance of good design in delivering good outcomes. Derek provided a very interesting presentation on the need to move away from design processes that just prioritise short-term capital expenditure, often to the detriment of running costs. "We need to think longer term, and not have capex alone dominating, especially if we truly want to be sustainable," he said, before going on to future gaze a little by discussing the uses of renewable energy, robotics and smart materials, possibly nanos in walls that can change the thermal properties of a building.

Case study
Gerry Demarco, a managing director at Morgan Stanley, was up next showing how to manage a massive changeover process, which has so far involved 3,500 servers being moved, 400 applications and 43 migration weekends in offices stretching from London, to Tokyo and New York. He asked for a show of hands of those who've undertaken a migration project or will do so shortly. As the majority of the 100 delegates raised their hands, Gerry nodded sagely and said: "Everyone, right. It's a common process, so let's do it well!" He went on to illustrate how his bank has handled its application migration programme highlighting the roundtable management procedures they put in place, involving preparation, processes and outputs, and the 'ten week wheel' they used to handle every move (see Passing Shot HERE for more on this, although not everything from his presentation is included in this standalone feature as Gerry's talk was confidential, with much of it only available to those attending on the day). The scale of the 'AppMig' programme, as Gerry dubbed it, is illustrated by 400 applications moving so far, 500 terabytes of SAN storage, and 1,000 old servers being decommissioned with half re-provisioned, not wasted. Ten things to do before you migrate, such as removing references to IP addresses, were also spotlighted. "I think our work has helped to extend the life of Morgan Stanley's existing facilities and fill up our new data centres efficiently," concluded Demarco.

Prior to lunch, Barry Maidment, product manager, Rittal, talked about how you can improve efficiency via hot/cold aisles, chillers and the deployment of technologies such as liquid cooling. He also talked about how 'green data centres' just used to be called good design, and the effective ban on any more data centres in and around London until after the 2012 Olympics as the games are expected to draw an additional 400MW of power.

Energy conservation was the touch-stone topic after lunch, with Hassan Moezzi, director, Future Facilities, explaining the physics of the data centre, its thermal properties and how to use this knowledge to get optimum performance. The use of CAD modelling techniques was covered, as was the need to increase power densities in the entire supply chain, deploy hot/cold aisle containment and so on. The effectiveness of using a virtual data centre to improve the performance of the real thing was illustrated by a video showing the modelling work Future Facilities did on Citi's European data centre in Frankfurt (a category winner at the FST Awards 2008).

Latency
An overview of the Chi-X multilateral trading facility (MTF) and why the data centre/latency is so important was provided next by Florian Miciu, chief technology officer, Chi-X Europe, who explained the requirement for fast trading services to enable clients to obtain arbitrage and the vital role of the data centre in delivering this. The data centre's part in ensuring there is enough capacity to trade, business continuity and so forth was also covered. Additionally, Florian talked about what a CTO, such as himself does, and touched on other relevant topics, such as proximity hosting. "The data centre is one of the most important elements in our infrastructure," he said. "A good one can reduce technology costs by a factor of 10-20 per cent." The location is also critical [vis-à-vis speed] and the stability of latency. "If it's 300 microseconds, it must always be 300, so that clients can design their algo trading systems accordingly."

Up next was what Ashley Davis described as an excellent presentation from Mark Lewis, senior director, marketing & alliances, EMEA, at Riverbed Technology who talked about how to optimise your wide area network (WAN) and the next phase of consolidation, desktop virtualisation and disaster recovery in the data centre. Mark claimed he could make your
WAN perform like a local area network and also talked about the 'swinging pendulum' - namely, the way mainframe and client computers dominated in the 1980s, before giving way to PCs, and how VDI and the cloud are bringing back clients.

The final presentation of the day was another case study from Graeme Buchanan, programme director, National Australia Group, who talked about the construction of NAG's new Tier 3+ Scottish data centre for Clydesdale and Yorkshire banks, which won the FST Award last year for best data centre project, and its lift and shift migration. Lots of tips were provided, such as how to persuade the board to sign-off the money, the need to talk to your SAN/connectivity supplier early to ensure a smooth changeover, and using the requirement to re-rack and re-cable as an opportunity to revolutionise layout. The bank virtualised its operation afterwards as well, rather than simultaneously, as they thought the move itself risky enough. "My background is finance and I wasn't initially keen to build a new data centre but once we looked at the case it made sense - there were no suitable co-location options in West Scotland for us and the efficiencies on offer were attractive." The bank used free cooling, raised floor grommets to keep cold air in the room, and rack blanking panels, among other techniques, to get a PUE of 1.7 versus an industry average of three. The latter technique alone can ensure a two per cent power saving, explained Buchanan, who also warned that big project risks still remain ahead of the migration and shouldn't be ignored just because your new facility is built. He concluded by offering tours of NAG's new facility to any other banks, as they benefitted from this kindness themselves.

Quite a few cards were swapped after Ashley Davis brought proceedings to an end by reminding people of the flow of the day, looking sequentially at the market, maintenance, design/modelling and the platform - i.e. servers, storage, and so on - all illustrated by regular case studies and comment from leading speakers from the industry. Some delegates no doubt arranged trips to go and see NAG's new data centre over a few beers as the day drew towards a close. As Morgan Stanley's Edmonds said: "There's a lot to think about in this virtual world." The DCEV Conference gave financial services professionals the time and space to do just that.

Share Story: