Happy April's Fools Day

Well, as you probably know, today is the day that the Conficker virus changes its mode of operation and starts checking with different servers on the Internet to see if there is any code for it to download. (If you want the full details have a look at this paper.) From what I've seen to date, news reports have been balanced on the whole, apart from the odd tabloid. So far, it would appear that nothing major has happened.

We'll wait and see. Those of us who thought we weren't being foolish might be made fools of yet.

B

Installation vs Implementation

I attended one of Joe Peppard's IT Management Masterclasses in the ICS last week. This one was entitled "Designed to Fail: Why IT Projects Underachieve And What To Do About It". One of the reasons projects fail to deliver what they promised is that they are over-sold to get them approved in the first place. Unsurprisingly, Joe did touch on this. However, for me the most interesting slide of the day was the benefits dependency network. Not new, but the first time I've seen it applied well to IT projects - with examples.

We know that IT alone cannot deliver results. Yes, you can install an ERP, CRM, ecommerce portal, collaboration, content management, or [insert name here] system. But if it is not used by the business people in a way that benefits them and the organisation, then the system has not been properly implemented and the project is a failure. In such a scenario, it's likely that the IT techies will say that the system (hardware and software) has been delivered, and it's not their fault that the business aren't using it.

I'm cautious when approached by a business manager seeking a magic bullet system to solve his or her problem(s). I start off by saying that it sounds like a good idea (if I genuinely think it is), explain that (typically) the hardware and software installation can be achieved reasonably easily, so we'd tease out the details of the implementation. For some, this is the first time they realise that the implementation will involve a lot more than hardware and software installation (not to mention configuration).

Joe's presentation can help illustrate the difference between installation and implementation. IT/IS - the installation - is on the left, with the investment objectives on the right, and the business changes - a key part of the implementation - in the middle. Have a look at it here. The pertinent slides start on page 22, work through the model and a couple of examples, and finish with a summary on page 39.

I find it illustrates very simply and vividly the difference between installation and implementation. Worth a look.

B

What are OLAP Cubes Anyway?

I got talking about Business Intelligence and OLAP cubes with a Business Analyst and an IT Manager recently. I explained to them in simple terms what OLAP cubes were all about.

But rather than re-inventing the wheel here, if you want to learn about business intelligence and OLAP cubes in particular, have a read of this relatively short and well written article on the subject.

B

Two Brians, Three Billion and BI

Brian Cowen and Brian Lenihan are making headlines for all the wrong reasons. Cowen, because of the over-reaction to the nude portraits story. Judging by public opinion, it would appear that the original story would have received less coverage if RTÉ weren't asked to apologise, and if the Gardaí didn't question Will Hanafin of Today FM.

The government should and do have more to worry about than a couple of paintings. The €3 billion fall in tax revenue - for example. The worst part of that story is that the incorrect figures were only projected in January.

One aspect of Business Intelligence software is predictive analytics which can be used to make forecasts based on historical data. Perhaps the Government could use some.

But to be fair, predictive analytics can only be as good as the data set available for analysis. My guess is that any data set available to the government is of no use to them. The data sets don't contain the patterns we're seeing today and so can't be used to make predictions to any realistic degree of accuracy. We're in unchartered water. One thing seems to be certain though: whatever figures the government come up with, the reality will be worse. At least for the foreseeable future.

B

Deep Purple

Fuel is almost constantly being added to the story that IBM is in talks to buy Sun, one of the latest being the Intel CEO weighing in on the IBM-Sun Talks. If it happens, some will no doubt be sad to see Sun go, and having been involved in the purchase and implementation of a small number of Sun boxes over the years, I've a bit of grá for the company myself.

But business is business. Sun suffered after the dot com bust, not least due to the shift to Linux on generic x86 hardware, which was cheaper than buying a SPARC based system to run Solaris - which was all the rage when there was plenty of VC cash being sloshed around and spent recklessly. Sun responded with OpenSolaris - but too late.

It's no surprise to read Jonathan Schwartz (Sun's CEO) blogging about their market potential, open source and their new cloud initiative in his recent post. Given IBM's relationship with open source and cloud computing, it'll be interesting to see what they do with Sun if they make the purchase.

Assuming the deal goes through, they'll have to have a project called Deep Purple. It's too good a naming opportunity to be missed.

B

Wales vs Ireland and Super Tuesday

Well, we won, but only just. Nevertheless, a win is a win, and we'll take it gladly. I expect the euphoria will die down fairly quickly, especially as "Super Tuesday" looms.

I sincerely hope that this "mini" budget will have a greater positive impact than the last "real" budget: it needs to. We need decisiveness from the government and a budget that they can stick to. Implementing budget changes and then backtracking on them is much worse than not making the changes in the first place - for so many reasons. This "mini" budget needs to be a first step along the path to economic recovery.

A win is a win? In sport, yes. But in this case, we need a decisive victory - not just a win.

B

Wales vs The Future of Ireland

Having heard the vox pops in the media during the week, apparently I was oblivious to the fact that the future of Ireland is in the hands of Dr. Declan Kidney and his team of celtic warriors today in their game against The Joneses in The Millennium Stadium. If we win¹, it's economic recovery, good summer weather and cheer all round. If we lose, it's donkeys instead of cars, pestilence, famine and emigrant boats (although where they'll be headed is an entirely different question).

My prediction? If ROG holds it together, (and I thnk he will, now that he's over his personal psychological barrier) - we'll win it - but only just. And the future of Ireland? Ahh sure, begorra - it'll be grand. Now, where's me leprechaun....

B

[1] "we" - even though most of us aren't actually playing

Aircraft Overhaul Solutions Ireland vs SR Technics

As you may have heard or read in the media, a company called Aircraft Overhaul Solutions Ireland Ltd. (AOS) was formed earlier this month with the intention of purchasing plant from SR Technics who are pulling out of Dublin. As an ex-Avionics Engineer who worked in Aer Lingus Maintenance and Engineering (M&E) and lived the ugly transition to TEAM Aer Lingus, I have a particular grá for what was Aer Lingus M&E. It's a shame that a large employer who essentially bought the assets of what was once a semi-state company is now pulling out leaving over 1,000 redundancies in its wake.

I sincerely wish Eamon Russell, Sean et. al. in AOS the very best of luck with their plans to buy the Dublin assets of SR Technics and establish as a viable aircraft maintenance company. Whether SR Technics will be willing to sell their assets to a start-up competitor is another issue.

No matter what happens, I wish all my ex-colleagues the best of luck for the future.

B

FreeNAS, Openfiler, RocketDivision and Microsoft's iSCSI initiator

I had occasion to set up a two node (Active/Passive) cluster using MSCS on W2k3 recently in a lab environment. I could have used ESX to host the nodes, but I needed it to be more portable so I set it up on VirtualPC instead - accepting the limitations that that imposed e.g. number of host CPUs, number of virtual networks.

Most cluster implementations need shared storage, and shared storage isn't cheap - not cheap enough for my lab anyway! So I had a poke around the various free offerings and opted for RocketDivision's StarWind iSCSI target software. StarWind is easy enough to get up and running, but I had problems with their iSCSI initiator so I used Microsoft's own iSCSI initiator instead (for W2k3 - it's OOB in W2k8). This solution works fine and I recommend it if you need to set up MSCS in a lab.

I also needed an NFSv3 NAS when prepping for the VCP, so I tried out Openfiler on an ESX VM. Openfiler installed OK, but I couldn't get the ESX host to connect to the NAS I configured. I didn't have time to troubleshoot the errors generated, so I fired up FreeNAS instead. FreeNAS installed fine and ESX connected to the NAS I configured first time. I found this document on FreeNAS Datastore Configuration useful. It's a bit dated, but you can still follow it on the current version of FreeNAS.

So, to sum up RocketDivision StarWind and Microsoft's iSCSI initiator for MSCS and FreeNAS for a NAS for ESX. Worked for me.

B

PS I could have used Windows Services for UNIX to provide me with NFS functionality, but I figured it was a good opportunity to step outside the Microsoft product suite

RSnake, Newspapers and Programming Errors

The CWE / SANS Top 25 Most Dangerous Programming Errors has been published. It reminded me of something I read recently along the lines of "show me a newspaper with no typos and I'll show you software with no bugs". Programmers are human, errors occur.

Robert "RSnake" Hansen touched on some of those top errors in his presentation at NITES recently. Having had the pleasure of chatting with Robert on his "everything is insecure" stance, I'd actually re-write that quotation as: "show me a newspaper with no typos and I'll show you software that has no known security flaws". Errors are fixed. But hackers are human. They'll think of new ways of breaking software. Humans are inventive after all.

Software development - like anything - has risks. Managing the risks is what's important. Adopt a layered approach. If you have anything to do with software development and don't know what those risks are, educate yourself.

B

An Investigation of Malware Protection in Irish SMEs

I've (finally!) got around to making the dissertation I wrote for my Masters in 2004/5 publicly available. The title is (as you might have guessed) "An Investigation of Malware Protection in Irish SMEs".

If you're interested, you can access it here.

The origins of the dissertation are rooted in the fact that there was much research done on malware protection in large international companies, but little or no research related to Irish SMEs. The situation has changed somewhat these day with the availability ISSA / UCD Irish Cybercrime Surveys.

I discovered that Irish SMEs were being negatively affected by malware infections. Best practice malware defences were not being implemented in part due to practical limitations such as lack of human resources and lack of knowledge. I suspect little has changed.

B

SharePoint Document Library or File Share?

I had a conversation the other day on SharePoint Document Libraries. The issue discussed was whether SharePoint Document Libraries should be used to replace traditional file shares. We concluded that unless there is a compelling need for version control or workflow, then SharePoint is overkill e.g. it has a significant systems overhead compared to a traditional file share and it introduces limitations which may not be fully understood by users in advance of an implementation.

Someone with a hammer thinks everything is a nail. IMHO: Don't use SharePoint Document Libraries unless you need to.

B

VMware Certified Professional

Well, I sat and passed the VMware Certified Professional Exam (VCP VI3) last week. It was tougher than I expected - which is fine as it increases the value of the VCP in my own mind. The exam also asks questions on VMware Server, so be sure you know the differences between VMware Server and VI at a minimum before you take the exam. And be sure to know your ESX partitions as well.

If anyone is thinking of taking the exam, you definitely need to get hands on experience with VMware Virtual Infrastructure (VI). There's a few ways to do this e.g. a) in your place of employment b) on your own physical kit or c) using ESX / VI on VMware Workstation.

a) In your place of employment
Poke around, have a detailed look at the different elements presented to you on screen. See what you can configure and where. Most importantly, if you're poking around a production environment, don't break anything ('Cancel' is useful).

b) On your own physical kit
Contrary to what some may think, ESX can and does run on desktop hardware. For starters, try Ultimate ESX Whitebox. That's where I started.
NB Be aware of potential gotchas e.g. I could boot ESX from the install CD on my kit, but couldn't install it. The IDE CD/DVD-ROM drive turned out to be the root cause of my problem. I swapped it for a SATA drive and all was well.

c) On VMware Workstation
Yes, you can run ESX on VMware Workstation. But it ain't the easiest route to go. For starters, look at 'VI in a box'

I opted for b) and c). It allowed me to do whatever I wanted without putting a production environment at risk. It worked well for me: I passed the exam. Mission accomplished.

B

PS Google is your friend, as is the VMware Community.