I've just spent 45 minutes on the phone with an expert from Sun talking about how I can squeeze even more performance out of the T1000 after my report, T1000 web performance results initial thoughts.
It was one of those conversations that was littered with more TLAs (three letter abbreviations) and expressions that would have sounded like Martian to the general populace. It was, despite this, phenomenally useful.
It also shows just how complex the optimization process can be; there are so many things to think about, from the performance of the network hardware, the driver that supports it, the OS stack that provides this interface on to the application, and of course the application itself. With so many things to consider, it's no wonder that it can take months to get the best out of a system.
I don't know about you, but I'm seeing an increasing quantity of spam coming through the email box at the moment. Despite my best efforts of updating custom spam filters and doing a lot of analysis to determine the trends so that I can write those rules, new spam keeps coming in quicker than I can handle it.
Today, though, I tried a different track. I've been using server-side filtering (through the Sieve tool offered by Cyrus, in turn through Mac OS X Server) for years to organize my folders and help me prioritize the email comes in. By placing the mail I want quicker access/responses to in specific folders, I can safely ignore the mailing lists, order confirmations and other non-priority mail.
Regular readers will know that I've spent the last year or so trying to reduce and migrate my server requirements and computing needs down to a select few machines, partially to reduce the amount of kit I have lying around to support my work, and partially to reduce the ever increasing electricity costs.
Over the last year, I've been able to simplify my equipment largely because of the move by companies to x86 hardware. Through Parallels on Mac OS X I can run pretty much every operating system I need to use.
I've decommissioned a number of servers, big and small, and now rely on just one machine - a 17" iMac - which handles my file storage (through a couple of big external drives), my email, web serving and other needs, all through OS X server. It may not be a huge super power house, but it is silent and perfect capable for the level I need internally.
Widgets, as demonstrated in Mac OS X's Dashboard and used frequently in many business reporting tools are a phenomenally useful way of viewing and interacting with select and specific pieces of information.
Soon you'll be able to get a widget display solution for the desk, in the form of the WidgetStation, a small device with one colour and one monochrome display for displaying a small discrete amount of information. Unfortunately it's not Wifi, hard wired only, but other companies are producing similar devices, like Chumby.
Well, it was kind of inevitable, but the UK has been hit by a major security leak in the form of a stolen laptop, as detailed at the BBC with Security raised over laptop theft.
Eleven million customer records have been stolen. It's only name, address and account details, but it is more than enough for a fraudulent individual to try and get money from the accounts. Rather frustratingly for those people who are customers of the Nationwide, the news of the laptop has only been released three months after the laptop was actually stolen.
Well, we've just laid our eyes on the Typhoon 600 series, which is loaded to the brim with Clovertown chips, reaching 256 gigaflops worth of processing power from a single outlet. Of course, the Typhoon isn't exactly going to crack the Top500 list of supercomputers, but it'll probably work much better than that those decked out quad-core getups we've seen thus far.
There's a great (but short) piece on the Top 10 data disasters revealed on the BBC website. As the article points out, the storage of critical data is becoming more important now that consumers store their photos, music and other files electronically.
Some of the stories (well, short summaries) here are disturbing. I can't imagine how you would accidentally run over a laptop, or decide that spraying WD-40 into your laptop was a good idea.
It highlights, however, some simple truths. Backups are vital. Having to use mine recently (and I'm really meticulous about my backups), even I found some faults and flaws in the methods used, and curse myself for not verifying the process sooner.
We had an interesting internal discussion a few weeks ago about the relative merits of dual monitors over big screens. While there were pros and cons on both sides of the argument, it's really down to a question of personal taste whether you prefer one model to the other.
The critical parameter though is the screen real estate - the actual number of pixels available, rather than how they are configured. It doesn't matter whether those pixels are spread across one screen or two. My two main machines - a 24" iMac (1920x1200) and a 17" MacBook Pro (1680x1050, plus an external 20" with 1600x1200) feel equally productive. On one, I have enough pixels to put many windows side by side. On the other, I spread them across both monitors.
In a new article in my System Administration Toolkit series I examine the role of swap space in your Unix /Linux system and how to configure and optimize your swap space usage. I also demonstrate a few tricks for helping you out when you run our of swap space unexpectedly.
Configure your swap space (including adding space in an emergency) to get the most out of your system. In this article, you’ll learn how to monitor your system to determine an effective swap space figure as well as examine methods for using swap space for more than just secondary random access memory (RAM).
Whether you are an established grid manager, just in the process of developing and designing your grid, or expecting to take over a grid environment soon, there are many different aspects you need to consider.
In Part 2 of my new series, written by Arun Chhatpar, we look at the security requirements and considerations in your grid.
Managing a grid involves many elements, from the network and hardware you use to deploy your grid to the security, job management, and job metrics and statistics generated during the execution of your grid, which enable you to more effectively manage the work. In this four-part “Managing a grid” series, we look at key elements of the grid management process, such as identifying hardware and network fundamentals and how that affects your grid process, and how to use metrics information as a scheduling, prediction, and expansion tool. Here in Part 2, we cover grid security and its importance in grid management.