Archive for June, 2007
June 29, 2007 at 1:52 pm · Filed under How IT works, Other things
http://www.guardian.co.uk/worldlatest/story/0,,-6745635,00.html
LONDON (AP) - British police defused a bomb found in a parked car in central London on Friday, and the new government called an emergency meeting of senior security chiefs to investigate what many feared could have been a planned terror attack.
A British security official told The Associated Press that the car was packed with explosives, gas canisters and nuts and bolts and would have caused “significant damage.”
He said there were similarities between the device and vehicle bombs used by insurgents in Iraq.
I raise this because on BBC Radio 4 this morning it mentioned that a number of companies/offices in that area had to invoke their disaster/contingency plans to deal with the various closed roads and disruption.
It’s always important to plan for emergencies, to understand what you expect of your teams to do, how business is set to continue etc.
June 28, 2007 at 10:54 pm · Filed under Other things, datacenter
http://www.onestopclick.com/news/Data-centre-to-flood-rescue_18194657.html or http://www.ehiprimarycare.com/news/item.cfm?ID=2823
UK data centre has ensured that GP practices in Lincolnshire and Yorkshire have been able to maintain their business continuity despite floods which have seen parts of England submerged under feet of water.
At least ten GP surgeries had PCs badly damaged by the floodwater, but thanks to the remote data storage policy of Computer Science Corporation (CSC), the medical practices should be able to continue without losing data.
Very cool, it’s nice to see the technologies being used to maintain data for remote access by those GPs that have been affected by the floods in the UK recently.
Anyway, an interesting article, the second link has more detailed info, a good read!
June 28, 2007 at 10:48 pm · Filed under Grid, Other things
http://www.sun.com/servers/hpc/SunConstellationPreview.jsp
The Sun Constellation System is the world’s first open Petascale computing environment combining ultra-dense high performance computing, networking, storage and software into an integrated system that delivers massive scalability, dramatically reduced complexity and breakthrough economics. The Sun Constellation System brings key innovations to market and is designed for radical simplicity that puts a new level of high-performance computing within reach for customers, whether they are pushing the limits for what is physically possible or if they are hungry for more performance from HPC solutions in production today.
Very cool, both IBM and Sun (in this article) are showing their Petascale computing solutions, it will be interesting to see what applications they can be put too. Check it out, an interesting read.
June 28, 2007 at 10:46 pm · Filed under Other things, guestblogger
IBM has devised a new Blue Gene supercomputer — the Blue Gene/P — that will be capable of processing more than 3 quadrillion operations a second, or 3 petaflops, a possible record. Blue Gene/P is designed to continuously operate at more than 1 petaflop in real-world situations.
Blue Gene/P marks a significant milestone in computing. Last November, the Blue Gene/L was ranked as the most powerful computer on the planet: it topped out at 280 teraflops, or 280 trillion operations a second during continuous operation.
Put another way, a Blue Gene/P operating at a petaflop is performing more operations than a 1.5-mile-high (2.4km) stack of laptops.
Very cool and very powerful, it’s always great to see what can be achieved with the technology, and this IBM supercomputer sounds fantastic, the zdnet article talks all about it, and the research.ibm.com site has great pictures of it.
June 28, 2007 at 10:37 pm · Filed under Other things, guestblogger
Got an email from a reader saying, have you read this? I think this would be an interesting post! It was an article about the HP and Microsoft partnership about Microsoft Windows Compute Cluster.
It’s great that people are reading the blog and thinking of articles that fellow readers might find of value. So if you do find anything on your google searches, your web travels and you think it would be of interest, email me and I’ll post it and refer to you if you like, by posting it under guest blogger.
To Ian, and Matt as ever, thank you for your emails, the articles you highlighted, thanks.
June 28, 2007 at 10:23 pm · Filed under Other things, blades, vmware
I had dinner with a friend who’s been writing some documentation for his companies projects to implement blade workstations, and desktop virtualization, as we had dinner, I took the opportunity to quiz him about them and this is the result:
What blades will you look at?
Well we’re going to test the HP and the IBM blade workstations, we’ve been playing the ClearCube ones they’re very good, but we have a purchasing agreement with HP, so that can sometimes sway things, we’ll be selecting blades with typically 2x Xeon processors, 4GB RAM and a 60GB disk, we’ll need to be able to drive at least two monitors up to four.
How are you going to virtualize the desktop?
We were thinking of either Citrix and Ardence where we might have a boot image which the client loads and runs against, or a VMWare virtualized desktop session, we’ll be trying both, I was thinking something like a DL585 as the hardware, SAN boot, SAN storage for the images, the main objective I’ve got is reducing the desktop disk size, we’re buying pcs with 80GB drives these days, we only need XP and bank applications so we’re looking to half that.
I remain a big fan from the centralization part, but I do wonder about the cost element, how have you found it?
Well have you thought about the desktop networking side? All the switches and routers I need in order to provide and maintain a production desktop network? When I virtualize the desktop everything’s on the central core - everything’s at the server side and basically you’re just talking to it, the network infrastructure between the client and the server gets simplified, that’s a hidden benefit.
On the support angle, it’s the general virtualization benefits, and the ability when a user complains that their desktop is slow, to take a pre-built image, point them at that, have them log in, they’ll pick up their applications/settings and start working on a new pc. Doing that in the physical world is expensive.
Do the traders actually need a blade workstation?
Not all of them did, but what we offer internally is a kind of Big Mac meal choice, you’re either a high performance user, or a standard user, standard users get a thin client that can handle one monitor, has a few usb ports and runs on our virtual infrastructure, it’s say £750 a year including support and the thin client. For the power user, you pay double that for the ‘instance’ the blade workstation, and you buy the monitors you need. The blade is dedicated to you, we have a few hot swap ones, and anyone can have it, they just need sign off. We’re anticipating that most people outside the trader arena will cope with the standard user option.
We’re opting for either blade or virtual to keep life simple, there should be a half way but we discussed it and wanted to keep it simple, we can offer more memory/disk on the blade pc or the standard option if needed.
How much do you think you’ll save?
Well I have to be careful what I say hear, but let’s just say that it’s a significant cost, the desktop support hardware costs should fall, the actual desktop support calls should also fall as we are able to better control the environment, coupled with the energy efficiency of the thin clients and the blade pcs, we can centrally control them, have the virtual instances paused overnight?
June 28, 2007 at 9:53 pm · Filed under How IT works
A lot of money is wasted in the enterprise through poor communication, through the detachment of the infrastructure teams, the evolution teams and the business/end users. That the evolution teams, the guys that work out what operating systems, what servers, what tools to use, might make a decision which causes disruption or has the opposite effect of the original intention.
I’ll give you an example, I was speaking to Mike the other week, he’s working for one of the British banks, and he’s been handling their Windows XP roll out. One of the problems he has is that the bank has many branches in the UK, the pc is there, but if the pc doesn’t work, it’s not the end of the world, the main work is provided by their citrix/thin client solutions, with that in mind, they operate on a three year or so replacement strategy. As a result the rules for desktops are set with a retail bank mind set, a Pentium 4 with 256mb ram and a 10gb hard drive is the ‘minimum’ supported spec.
The problem is, the business units in the investment bank subsidiary sometimes use these ‘rules’, these guidelines to control their costs, that a document somewhere says a Pentium 4 is suitable to run XP, means that he sometimes gets requests to rebuild a pc which is nearing it’s end of life in an effort to avoid the £150 cost of purchasing a new pc, to keep it going that bit longer. You can completely understand this, in the branch with three people, where the pc is used once or twice a day, but not so much in London, where it’s the centre of the users’ day, their universe.
With that in mind in the banking sector, in the enterprise shouldn’t there be two strategies? Ideally you’d have one, the high availability, user focused solution where money is secondary to the delivery (not wasting money, just continually investing and evolving the IT for your business). However, I suspect this is more a want than something that will happen. So back to my original point, can I have two strategies please, for my two different business lines.
I’ll have one for the retail bank, a locked down robust windows platform where changes to the windows platform are managed, controlled and where the hardware standards change infrequently. I’ll also have the investment bank equivalent, where your pc, your server is (either virtual) or depreciated over a maximum of three years, where at the end of it, you either buy a new box or as the box ages, your support costs rise, where there is no incentive to keep a computer over the long term. That way I avoid the following:
- Windows migration projects - they are expensive and disruptive, not to mention quite emotional - but why should I buy a new box? My NT4 box runs fine
- Conversations about upgrades - the memory costs how much - my box would typically be nearing replacement by the time it needed upgraded
- Make the server a commodity - it’s a component of the infrastructure - not the center of my existence -but I need my Compaq…
- My application code needs to continually evolve - avoiding those hard coded hostname/ip addresses in the application code
- Avoid security compliance issues - again the box gets rebuilt/replaced every 3 years, at which point we can fix/evolve security with the new server
- Avoid the asset problems - as the name changes, as the box swaps hands - “LONS19872? that’s an transversal system, not mine..”
- Being stuck on set hardware/infrastructure - that application which is only validated to run on Dec Alpha kit on VMS version…. sigh…
Interestingly virtualization might resolve some of these key issues, as I abstract the operating system from the hardware, and as we reach the point where for example the application can run on just an ESX instance without Linux or Windows, this will be even more the case.
In the meantime, can I have multiple strategies please, all with a focus on delivery, common sense, long term cost avoidance and consistency in user experience.
June 28, 2007 at 4:44 pm · Filed under Abn Amro merger
http://investing.reuters.co.uk/news/articleinvesting.aspx?type=mergersNews&storyID=2007-06-28T143048Z_01_L28678269_RTRIDST_0_ABN-BARCLAYS-JOBS.XMLÂ
LONDON, June 28 (Reuters) - British bank Barclays Plc (BARC.L: Quote, Profile , Research) said it plans to avoid compulsory redundancies where possible for its employees and for staff at its Dutch takeover target ABN AMRO Holding NV (AAH.AS: Quote, Profile , Research).
Barclays said on Thursday it wanted to provide clarity and assurances to staff following meetings with employee groups, including trade unions and works councils.
Barclays has said 23,600 jobs will be cut if it buys ABN, including 10,800 positions moved to low-cost locations. The total would represent almost 11 percent of the banks’ combined workforce of 217,000.
The deal continues, I wonder what reception this will receive.
June 28, 2007 at 3:47 pm · Filed under Grid, Other things
http://www.it-quants.com/
This is the “unofficial” Sophis user group web site. By unofficial, it means that we are not sponsorised by Sophis, even if their members are authorized to take a look on the content of this site. We have nothing to do with events, restricted to customers, and organised by this company too.
We provide meetings and discussion forums in a neutral environment where all users, clients and contractors, can network, exchange experiences, and receive presentations from selected relevant companies. We do not pass membership details to anyone outside our organisation. We are an independent organisation, operated and managed for the benefit of our members.
How cool is this? Well done to the group, they’re using the technology as an enabler, it’s a great way of handling support issues, getting feedback on the solutions/software applications as well as casually seeing what issues people are having so you can be proactive with your user community, do check it out.
Sophis provide risk and value products, they have a vast range of experience in these fields and I know quite a few banks use Sophis on and off the grid, check out their site www.sophis.net.
June 28, 2007 at 3:27 pm · Filed under Other things, blades
The following is a list of blade server manufacturers and what I see as their potential benefits, in the interests of being impartial, if I forget anyone, do let me know, it certainly wasn’t intentional, and it’s simply to illustrate what vendors are out their, that there are many options all with different products that might be right for you and your business.
- HP - I love the HP driver pack, the smartstart process, Systems Insight Manager is brilliant, and their documentation is very cool.
- IBM - the range of enclosures is simply stunning as is the connectivity and blade processor combinations avaialble, the different models/operating systems can all be managed in Director easily.
- Dell - does exactly what it says on the tin, driver support is good, expandability is fine and relatively inexpensive.
- Sun - their new range of blades, the x86 and sparc powered ones would be an excellent consolidation/virtualization platform and the drive support is also very cool.
- Hitachi - couple them up with their virtualization solution, and their storage and you have a rather good virtualization solution. I’d need to properly try their virtualization software though to be sure.
- Fujitsu - an excellent alternative to the others with great connectivity.
- Egenera - one of the early blade enterprises which offers a very cool virtualization solution, they continue to evolve their blades and their solution concept.
- Rackable - DC option remains fantastic, can run quite hot if necessary, the connectivity options you’d expect as standard.
- Verari - very interesting blade servers with energy efficiency being an important part of the solution, very cool.
- Panta Systems - an Oracle ‘in a box’ solution using blades, it does sound fantastic and ideal for your development or production oracle database systems.
- NEC blades - 6 blades in a 4u rack solution, PCI-X slot support, and hot plug drives, very cool.
Next entries »
