Bladewatch.com

The inquirer.net

USING IMPECCABLE if slightly banal logic, IDC has declared that the drop in worldwide server shipments resulted in an across the board revenue slump for server makers in the first quarter.

The market research outfit reckons server unit shipments dropped off by a whopping 26.5 per cent year-on-year in Q109 to just 1.49 million units, the steepest slump in half a decade. Similarly, revenue was down 24.5 percent to a paltry $9.9 billion in the first three months of the year.

X86 servers purportedly felt the blow harder than lower-end Unix servers, with x86 revenue falling 28.8 percent to $5.1 billion, whilst non-x86 servers fell by 19.4 percent to $4.8 billion. This, IDC posited, was probably due to the fact Unix OSes deal mainly with mission-critical workloads, rather than faffing about with more generic tasks like e-mail, print and web serving which can easily be virtualised anyway.

With the banking sector having rebooted earlier on this year, economic conditions not necessarily being wonderful and that dangerous concept of many finding that their existing servers are good enough, there has been a report about server shipments falling. An interesting read, do check it out.

We’ll have to see, I think we’ll see this change in the near future, the new Intel processor has some impressive performance and energy efficient statistics, (just compare a DL380 G4 and G6 for energy efficiency and you’ll see what I mean), combine that with Windows 2008, further movements into virtualization and I think we’ll see sales improve. At the same time, the banks that have been merging, getting bought out or selling elements of their business will all need integrated to their new homes, they’ll need new servers bought to our standards, coupled with on going business requirements as well those white labelling products and service we see in the finance, trading and insurance markets kicking off.

Share and Enjoy

There was an interesting debate that popped up online regarding VMware’s policy for VMworld 2009, in reference to terms and conditions for vendors attending etc, anyway, they have clarified their standpoint here.

Share and Enjoy

The more barriers to entry with the technology, the more people will think about it before deploying it.

We need to remove the barriers to entry with the technology, not just in terms of how difficult it is to use, but also:

• How it will benefit my business operationally or technically – it’ll make it faster
• What I need to consider if thinking about adopting the technology
• What process steps, go-live requirements or issues others have had with the platform
• What sites/blogs/documentation or whitepapers there are to make deployment, support and evolution easier

The more we surround the product with community, with dialog, debate and support, the more we can achieve from the product, the more we can manage failure, manage expectation and get genuine free feedback for product evolution or improvement.

How wonderful would it have been had there been a blog saying, if you’ve got a problem with your HP P-class enclosure, take the backplane bit off for 5 minutes, plug it back in and that will reset the enclosure’s ILO connections and should fix that red light issue. The first line stuff, the easy stuff, with obviously the generic, be careful what you’re doing clauses, need to be more accessible to reduce vendor support costs and improve knowledge of the platform or product, in doing so, changing the situation from, that’s a hardware thing to a, oh pop down press these two tabs and then if it doesn’t work call support.

Share and Enjoy

Information is indeed power, but sharing knowledge earns revenue.

I was having a chat with a colleague the other day, he was relaying the infamous message regarding information being power (in relation to being a windows server guy). He’s right for the short term, it creates opportunities for roles and contracts in the market. The interesting thing is though, the technology continues to improve, automated scripts, automated functions are making it easier, coupled with remote working, the number of organizations offering MCSE and certifications increasing the supply of ready skilled engineers.

Vendors and support staff alike need to recognize that information (knowing how to fix it, how it works, what might go wrong is power), but you know what, if you want me to buy-in, I want some information. Knowledge is empowerment, knowledge reduces your support costs, enables me to be more successful with the technology, reducing the chances of me saying … is rubbish, because I can’t get virtualization to work in my business either as a result of cost, process or technology.

With this in mind, the more how to’s, the more best practice, and suggestions of how we got it to work that we share, that the vendors publish, the easier and more prevelant the technology can become.

Share and Enjoy

Check out what Intel have been saying about their new 8 core processor, it’s an interesting read.

The article outlines the benefits in summary below:

The Nehalem-EX Advantage

• Intel Nehalem Architecture built on Intel’s unique 45nm high-k metal gate technology process
• Up to 8 cores per processor
• Up to 16 threads per processor with Intel® Hyper-threading
• Scalability up to eight sockets via Quick Path Interconnects and greater with third-party node controllers
• QuickPath Architecture with four high-bandwidth links
• 24MB of shared cache
• Integrated memory controllers
• Intel Turbo Boost Technology
• Intel scalable memory buffer and scalable memory interconnects
• Up to 9x the memory bandwidth of previous generation
• Support for up to 16 memory slots per processor socket
• Advanced RAS capabilities including MCA Recovery
• 2.3 billion transistors

Share and Enjoy

Check out this article talking about SGI deploying their modular data center solution, it does look cool, anything we can do to aid meet business requirements within the operational or financial constraints in business has to be a good thing, I’m off to read up more.

Share and Enjoy

I have always loved history. At university it was between history and IT, I went with IT and resulted in my current status. Anyway, this url covers a topic that I wanted to talk about:

“For want of a shoe the horse was lost.”

Why?

I had an interview with one of the managers from a financial organization which is undergoing organizational change, (many of the banks are, as a result of mergers and acquisitions or costs cutting), as part of this he told me of some challenges facing him.

I run the database team, we install the database products and configure the systems ready so that we can migrate the application databases to the new servers. There are a few hundred servers being installed, moved and rebuilt as part of the re-organization, however the standards have not been agreed and set.

One enterprise standards support Ingres, Oracle and SQL, the other supports only Oracle and SQL. (This issue is across the IT department, middleware and the Windows/Unix guys all have different standards of layered products or configurations), resulting in two choices:

1. We either delay work until standards are set and agreed.
2. We simply continue business as usual.

This results in an issue, we either get accused of delaying projects, or we face the conversations later when we state that an application is out of support as it is using an incorrect database platform.

For example, we deployed an Ingres database server last week, IT has charged them two days time to configure the database, migrate the data and set it up. If we no longer support Ingres, the business will get told you need to migrate to Oracle, please change your application code, please purchase a new server, database and middleware, then allocate a few days time for us to migrate to the new server with Oracle.

At the same time, I need to bring all SQL servers to the same service pack and make sure they are all patched, but since we haven’t signed off which database platform we’re keeping, I’m reluctant to dedicate an engineer or two to upgrading the database servers, this of course means we’re moving out of support on our SQL platforms.

The lack of a binary statement, of standards therefore is going to result in duplication of effort long term, unnecessary cost and that challenge of ‘staying still’. The teams wont want to make significant changes to the infrastructure or service improvement plans until they understand the strategy, the direction we’re going on, in the meantime the IT function therefore becomes at risk of standing still and not continually evolving not just their offering, but the platform.

Interestingly it’s not a technical problem. It’s a managerial resourcing and strategy problem. As a result of IT and management having not agreed or published the basics, we undermine the business as usual and project work, the vehicles we use to deliver.

The organization begins to stand still – I can’t make decisions because I don’t know where I will be in 6 months, I wont make the effort on that because we might not support that in future. In these situations we need:

• Guidance – binary in nature – yes/no answers to what’s in scope, what’s supported and not
• A focus on business as usual – evolution and improvement of the platform continues – never stand still
• Isolation of the business as usual/project from organizational changes (as best as possible)
  • Where we are, where we are going too – the quicker the announcements are made the better
  • What core layered components, database and operating system are supported
  • What business as usual projects continue -  service pack/layered component upgrades etc

Share and Enjoy

An interesting read talking about x86 being appropriate for specific mission critical applications or services, do check it out.

We need to focus on the positives, discuss the benefits and how we can make it work for our organization – would we use x86 virtual platform a front office retail function, possibly not, but for that market data platform it might be equally ideal.

Choosing the right platform for the business function will always be a challenge, we’ve made big progress in the virtualization platform space recently and long may it continue, addressing specific business and user requirements, at the same time, we need to move towards more standards, configuration and best practice models – remind the user community that the virtual platform needs secured as you would the Windows/Linux installation on the virtual machine.

There’s an interesting thing happening in many respects, like the car industry, we’re erasing the middle ground long term, moving towards either the industrial strength or the the super cheap, disposable short term platform which suits us now and maybe for the next two years when I’ll bin it and replace it. This in terms of energy efficiency, data center and the hardware support costs. Is the 1u or 3u server, the blade combined with grid and virtualization the next generation supercomputer or Enterprise platform?

There will always be demand for the industrial strength ‘Non-Stop’ high end solutions, the last forever ones, for those specific government or tier 1 market applications, but I wonder as the x86 platform develops, as what we can achieve with it in partnership with virtualization and grid, if that can’t equally fulfill a similar role at reduced cost?

Is it industrial strength I need or something that works most of the time when I need it? How much downtime can i accept and in a virtual world with fail-over, with load balancing is it always online or just many cheapy computers that can share the load, manage end user expectations and service?

What investment to risk ratio is acceptable to my business, my customers? For those free services will the end users care (or notice) if the system is ‘down’? We need to define the difference between it being ‘slow’ (which we could suggest is client/web based) and the unavailable state, the site/application displaying the ‘sorry we’re broken page’.

Share and Enjoy

The Inquirer is talking about the New Zealand government’s deal with Microsoft and how the government is wondering if it can achieve similar results or enhanced value from open source. Certainly using open source software can help reduce your licensing costs and be a more efficient way of providing the IT services you need in your business. My only hesitation with this is the support model, by that I mean, if you’re buying Microsoft Office for example in bulk with Windows desktop, the actual licensing costs can be relatively affordable, and there are many that can support Office and the Windows desktop. I can sense so many comments saying but Linux or Mac just works, that you can use Open Office and indeed you can, it’s all about getting the balance right.

We need to invest in the right way, obtain the maximum benefit of switching to open source not do switch for the sake of switching.

For me, the core things in an organization I would look at are:

• Open Office instead of Word/Excel where appropriate
  • Users applications supplied on demand via Citrix or the equivalent
  • You only get the applications you’ve requested or need
• MySQL for the databases – is there a reason we use SQL/Oracle or Ingres?
• Dare I write this – googlemail? For the SME it’s much easier and cheaper than having a firewall, email server and guys looking after it – buy it in.

At the same time we need to think of the billing, the cost of conversion and the marginal cost in migrating the user from the desktop. By that I mean if I can buy a users’ pc function and support of that user for a fixed cost, is that not cheaper than putting everything in Citrix, of virtualizing the applications, of setting up active directory and everything else. In essence is it not billing and cross charging that’s the problem, when I know it’s a fixed cost for the desktop, why invest in moving the user to the server, the data center?

Share and Enjoy

Microsoft Vista Service Pack 2 has been released, I’m off to install it on my desktop, check out this article or the Microsoft Technet article for more information.

Share and Enjoy