Martin MC Brown's Most Recent Posts

I was sat in the break area of a company a couple of weeks ago, minding my own business while waiting for a meeting, when I overheard a water-cooler style discussion at another table about cloud computing. The main topic of the discussion was the expectation for many companies and individuals that while many are using cloud computing systems, like Amazon's EC2, but that longer term people want to be able to mix and match their cloud computing environments.

I've long been a proponent of better security on your computer, and educating all users about the dangers of clicking on things they shouldn't.

Recently, I've noticed that certain sites and searches are returning with notes and even interstitial pages warning me about clicking in to a page that may be dodgy.

I've spoken before about my desire to try some different development ideas with the Arduino

Over the festive break there were quite a few announcements related to Amazon's EC2 service.

We just had our boiler serviced yesterday, and I noticed that the engineer has gone back to filling out a printed paper form with a traditional form of instrument called a pen.

Last week I stated that Employees cause most corporate data loss, with a specific nod to the problem of physical access to machines.

Another problem that has occurred to me recently is the issue of mobile devices. Although a lot of laptops go missing (10,000 according to a survey by the TSA), I suspect a significantly higher proportion of mobile phones and PDAs go missing. At a recent conference, three iPhones went missing within the first few days.

Here is a really worrying bill being pushed through the UK parliament which will give the government the right to keep a track of all our web views and emails.

One of the terms associated with cloud computing at the moment is Cloudbursting.

Using current cloud computing models for handling all of your applications and traffic can be costly, and using it for handling the peaks in your usage patterns can also be fraught with problems because of the time required to get the system set up and seeded with the information. Ramping up in the cloud is difficult since it may take minutes to fire up more hosts but your peak requirement may last less time than that.

There's a brilliant article on how it is mother nature that bringing down many data centers rather than our more familiar types of disaster.

It goes on to look at some other types of problem that you may be more familiar with, but the non-human, and particularly animal interventions make for some interesting reading.

I remember describing many years ago about the issues of security in the typical data center (or, as we called them back then, server room). Talking to a room of IT managers with their own facilities, more than half were stunned when I talked about physical security as one of the most important aspects of any security system.

A nice white paper on the differences between different large scale computing technology is available over at GridBus, The Grid Computing and Distributed Systems (GRIDS) Laboratory at the University of Melbourne.

The white paper provides a lot of information, including some more detailed information on the differences between the different massively-scalable computing systems, and how the current IT climate is ultimately proceeding towards the cloud computing infrastructure.

One of the problems that cloud computing is trying to solve is the issue of dealing with capacity planning for companies and the services that they offer. Current datacenters for individual companies, and where relevant, for entire websites, are designed to cope with a particular peak load.

There are lots of discussions going on right now about cloud computing is, I even touched on at least some aspects of that last week. This story at Slashdot touched on the issue, and the Cloud Computing GoogleGroup is full of the discussion. But I feel there isn't enough definition of what it isn't. So here's a list of what I think:

So far, cloud computing has been a solely Unix, and more frequently Linux, affair. There are some good reasons for this, generally Linux is easy to deploy, easy to copy and quick to update and reconfigure, all vital capabilities when you want to quickly clone and duplicate your host environments across tens or hundreds of hosts.

I'm late following up on my previous X4150 thoughts - I can blame conferences for that at least - but I shouldn't shirk my responsibilities.

It is very difficult, from a buyers perspective, to decide what that sweet spot is in terms of buying hardware. The current computer market, and more importantly the current computer climate, has completely changed the way that we look at buying computers.