Well what I do is write my blogs when I get an inspiration, this most often is not while I am in the office. Most of it is while I am travelling in trains, planes, automobiles. So 3 blogs today, however, they were stored for weeks and months when I wrote the draft in some airplane or remote location.

3 blogs in a day, I am catching up to my lost quota of 1 blog per week. I still have 10 saved up over the months.

The marketplace and hollywood have very restrictive views and impressions of the mainframe and many believe that IBM are the only ones that can supply such a thing. Well in reality the mainframe does not exist only in the minds of marketing and fiction. As many suppliers nowadays have equivalent and/or superior mainframe systems to IBM's.

Everyone who has not worked with mainframes has this impression; they are large, expensive, fast, complex, special, adaptable, reliable.

Everyone who has worked with mainframes knows the following; they are large, expensive but will be discounted on the hardware to make them look competitive, but software charges and maintenance will be expensive, they are not fast anymore or no faster than any other large system, you can only run zOS (MVS), they are not complex just totally proprietary. Not really adaptable, everything needs to run on the zOS instruction set need hypervisors (read overhead) to do conversions to run other interesting OSes like Linux, really good for the old stuff like CICS and old batch programs that no-one understands as the people who wrote the retired. They are special as a whole load of emotional baggage goes with them, companies cannot migrate from them as no-one understand the apps, so companies are held hostage, costs are often hidden in various ways (s/w, maint, service) so you can never get to the bottom of the TCO. As for reliability no different than any other major large system where you have good change control. Mainframes have good reliability records as no-one is allowed to change them. If you have other high end systems that you do not change they will be as reliable as a "mainframe". The mainframe hardware is not special compared to other high end systems. Historically they were better about 10yrs ago, but this myth continues.

So we either stop calling the z10 server a mainframe or we call every high-end server a mainframe. Sun and IBM high end Unix mainframes (servers), (not sure about HP, SGI etc) high end servers have better or equivalent features to the z10 servers.

Mainframe acid test, ask IBM what is the best plaform for every application that you want to run, what is the difference between their Linux actually Red Hat, AIX and zOS offerings. Is AIX inferior to zOS (aka MVS), is Red Hat not as good as AIX. Is Red had better than zOS. When should we use what.

Can I run Red Hat on a p-series
Can I run AIX on a z10 mainframe server (apparently it uses the same CPUs)
Can I run zOS on a p-series.

I get really confused with the positioning and confusion of AIX, Red Hat and z/OS. When do I do what with what. Or do I just do everything with everything and make a real mess. Good fun for the techies.

So with HP recently they came up with some logical partitioning with something called Dynamic LPARs, well that is old established technology, nothing new here. Available from others and established technology for a long time. However, HP engineering was always good, HP used to be a engineering company initially. So their high end UNIX servers can also be looked upon as mainframes.

Now Sun M-series servers have instruction retry, whole memory DIMM mirroring, crossbar ECC, more dynamic replacement of CPU's, RAM than any other supplier while the system is running and more. So you not get get more reliable than that and mainframes may not even have the reliability of some high end Unix servers.

Mainframe has been around since 1960's so the name is familiar but missused. Like many things in life, understanding, education and knowledge helps dispell these myths when once upon a time long ago they were something special. Now I like the mainframe, love to tune the assembler code, patch and zap modules, good for your old apps that no-one understands and your company is too scared to migrate to a new lower cost system. But as with everything technology moves forward and other high end systems especially large Unix servers are just as good or better and can be understood by more people.

We had a customer using an application on a customer platform, the supplier wanted funding from the customer to re-write the application to another Unix. Sun recommended to use Dtrace to find the areas that could be improved and help the ISV move from Solaris 8 to Solaris 10. However, the application supplier said that it would take too much work/money etc, it would cost less to re-write to another unix.

However, the OEM wanted the binary compatability value offered by Sun and to exploit Sun's multi-core chip technologies. So when the customer officially requested that the ISV work with Sun to determine if the application could be moved to Solaris 10, we discovered some interesting hidden agendas. Sun together with the supplier very quickly found areas in the application code for improvement at a fraction of the time and cost of a port to another platform. Application performance was improved by an order of magnitude, 10 x faster and qualification onto Solaris 10 was not a problem. Costs of taking the Solaris 8 binary compatible code, tuning it with Dtrace and moving to Solaris 10 were a fraction of re-writing to another platform. So beware, people may want funding to move applications to platforms, then ask for more money later to support a myriad of kernels. However, with Dtrace they can improve performance dramatically and shorten the time to qualify Solaris 8 & 9 apps onto Solaris 10. Buyer beware, the reason for not moving to the latest release of Solaris may not be technical.

By the way we also consolidated several servers runing this application onto one M5000 using Solaris containers to run many virtualised instances of the application. Ended up with nearly 100's of applications running on one server. The customer could now reduce the number or racks that they needed to.

Carzy world, moving applications from Solaris 8 to Solaris 10 is easy and you can consolidate your servers at the same time. Madness where will it all end. One big computer.

Make sure that all ISV's and application providers know the benefits of a faster time to market and reduced development costs that Solaris and Dtrace can bring them. Costs too much to qualify on Solaris 10, what a red herring. Make your application faster with minimum investment, use those Dtrace features. It is the developers best friend, I wish I had such a tool when I was a programmer.

The moral of the story is; All end users, ask your application suppliers or Sun to validate application qualification to the latest release of Solaris with Dtrace.

Why does this remind me of some builders when you are renovating your house. When they say "that will cost you", in reality it is a much simpler job. Maybe that is why the DIY business is so profitable.

While looking into all the mainframe claims we (Jeff Savitt) discovered, lots of incorrect comparisons that were used when comparing new mainframe systems to 2-3 year old Unix systems (new with old comparison). Well a colleague in Germany has just busted another rather silly comparison, albeit for a new Unix servers to old Unix comparison, you can see the write up here.

Now I speculated that the equivalent amount of rack space of Sun Niagara servers e.g. T5220 would be able to successfully consolidate (1250 per rack) more servers than a mainframe, note the term successfully. I remember a while ago some very large mainframe Unix/linux consolidations failed.

This article is well worth looking into, the power and cooling characteristics of the T5220 will definitely have an competitive edge over the lifetime of the mainframe especially once you take into account the cooling and power requirements of the mainframe. Plus you can start small with the 2U T5220, and build in low cost pieces with one small T5220 at a time. With the big iron mainframe, cost of entry and cost of exit is very much more expensive.

Anyway thanks to Joerg in Sun Germany for another very good and simple explanation of more realistic and sensible comparisons.

Many people think that technology is not important and that it is a commodity. Well there are two parts to this, those that want to cut costs and treat IT as a cost center. This article is for you. For those that want to use IT to beat the competition, which is what computers were designed to do for companies in the first place, that blog will come later. However, recently this very good article appeared. Here is and example where tehcnology is important as it can save you money, power, cooling and management costs, while improving security.

Standard office over last 15yrs with PC's (last century computing model).

Desktop PC on or under each desk.
Each has a transformer/power unit built in, normally between 300-800 Watts - which means lots of power required.
Each has a hard disk built in - which means lots of local data to back up, and a security issue to protect.
Each has a CD, DVD or maybe a floppy disk built in - which means even more easier to steal data.
Many deskstops will be different thus requiring different drivers/maint fixes etc.
PC's are left switched on overnight - which means a waste of company money/power (even when on standby).
Each PC has it's own CPU, many systems in your company will mean many different speed and type of CPU, more complexity and management, plus wasted rsesource as they are so underutilized.
Each PC has it's own operating system - which means a lot of complex managment required to maintain and update (means more costs).
If you have a laptop, you need the network to do most of your work, but you better make sure you do not loose your laptop and it's data.

Mobile workers may need laptops and that is OK, but if you count the people in your company who always sit at the same desk everyday, those are the ones who do not need a company laptop.

Often you need many admins to manage these, if a desktop PC fails the time taken to replace and restore all data is hours or days.

I base my security theft assumptions on the fact that most security breaches and losses of data are due to internal process and people.

Thin Client or a Sun Ray (true thin client with no local data) which can present (run) Windows apps. This is what I call the new millenium computing model. Big picture so that you can see there are no tricks here.

Small unit size of a large pocket book on desk.
Small transformer up to 100W (actually 5-10W usage. Thus save you r power bill by 90% straight away. ROI case solved.
No hard disk - whcih means no local data to backup, no local data to steal. Security problem solved.
No CD, DVD or floppy - so nothing to steal, no security required to manage. Security solved, again.
No local drivers required as the SunRay thin client has a network card and graphics chip.
There is no local CPU , fan cooling or anything like that, much less power usage even when left switched on.
No local operating system. Thus upgrade & maintenance costs reduced by orders of magnitude.
SunRays come with Wifi connections, so if the network goes down it is easy to get another connection. If a disaster occurs it is easy to move to another building which has the half of the data and work can start very quickly from the point where you left your work. All you need is your ID card to sign in, or your password.
In a disaster we do not have to restore 1,000 of PC's but just one server, ideally the backup server has all the data in another site already and work continues. No great drama or crises here.

Sunrays are not for mobile workers. How many of us are mobile anyway.

If the SunRay fails, you can replace in minutes or hours, no data to restore, as there is no data locally. So you can use one of these SunRay thin clients and keep your present screens or buy whicever you want, the SunRay uses standard screens. Big pictures so that you can really check it out.

This needs 1 x power connection, 1 x network connection and 1 x connection to the screen, mouse and keyboard, if all goes wrong it takes 2 minutes to replace. Uses about 4Watts of power and has no moving parts and totally silent. It is the height of a paperback book and you can see a credit card size ID card that you can insert to identify yourself to any Sunray in your network/company. As you can see you can have microphones and speakers connected if you really want to. You can take it out rush upstairs and put it in your bosses Sunray and he can see exactly what you are working on. Great when people are changing over between presentations.

So technology is not important, one requires lots of maintenance, power and takes a long to replace if broken, plus horendous security issues. The other technology solves all those business issues.

Have a look here to see how you can get a competitive edge.

I do not want people to get the wrong impression, I had some very good times working with the mainframe and I learnt a lot. However, time moves on and I moved with the technology developments to other platforms.

This is very similar to my favourite old car, that was a Saab 9000, 2.3turbo, Anniversary Edition. It had leather seats, wooden fascia, quiet, fast and safe.

I often would still like to have it but for todays requirements it uses too much fuel and would probably start to get too expensive to run. Since that car I have had a Toyota, Saab 9-5 and an Audi. I have choice of the open marketplace, if I stick to the old Saab 9000 I am locked in. Which is never a good thing.

So bottom line, we look at the mainframe with good memories but we need to move forward and adapt to new technologies. We need to let go and adapt. We cannot tow around our old cars with us forever.

The other argument is that that the mainframe is adapting like the Saab 9000 to 9-5, as we get new models. But the new car models do not need to pull or tow along the old models. The problem is that z/OS which comes from ESA, MVS, MVT just keeps on getting add-ons. But it is still has to support all the old apps, mainly based on a batch processing system (which no-one can understand anymore), with some bits to make it run OLTP, Linux and all sorts of goodies (read complexity). Also the CPU's still need to run the S/380, S/390, ESA/390, z/9, z/10 and more recent version of this instruction set, of which there are 100's of old instructions. This is why it is called a CISC processor, Complex Instruction set computer. In comparison to simpler SPARC and Power RISC Reduced Instruction set computers. So we have one hell of a level of complexity with mainframes running complex instructions on RISC processors or derivatives thereof. Now this is interesting and fun for the geeks, but not really for business. As all programs/applications have to go through many levels of instruction abstraction/translation etc. Basically mainframes due to their history have immense complexity. Complexity means costly management and difficult problem determination.

I would have loved to have stayed with the mainframe all my career, it was fun, I was quite good at it, people I worked with were great, but many others also moved to UNIX and open systems. Back in the 1980's I actually believed that I would be working with the mainframe for the rest of my career, but then I was young and there was only IBM and DEC. Chips also changed from Bipolar to CMOS, which made it easier for competition. Now the old heatsink/towers on Bipolar chips look not too dissimilar from all the heatsinks we now need on our PC's sad, that is why Sun coolthreads CPUs are cool. However, technology and the world moved on and the mainframe tries to re-invent itself. While carrying all it's historic baggage, which is some weight that we need to burden ourselves with. So like an old shoe, you can replace the soles so many times until you just have to get a new pair. Sometimes we have to try a new pair of shoes or a new design.

IBM compares new z10 mainframe with 3-4 yr old Sun servers and makes their own old mainframes look slower. Now I worked with the mainframe for over 10 years, COBOL and CICS and all thoses thing were great. However, these type of comparisons devalue the mainframe. We need to compare new products from one supplier (IBM) with the equivalent new and current products from competitive suppliers, Sun etc. Quote from IT Jungle article "Last week, with the launch of the System z10 Enterprise Class mainframe, the mainframe is as close to parity with big Unix and proprietary RISC machines as it has been in a long, long time." But can IBM keep this up, the mainframe lost competitiveness about 10yrs ago and the UNIX mainframes will not stand still.

Not sure about the proprietaty claim either, SPARC CPU designs are open source and have been for over 10 yrs, see OpenSparc. Also Fujitsu use the same CPU's. So Sun has other companies using their Chips, I do not know of another company using the IBM chips or IBM CPU designs being open source. I am afraid that Sun is not proprietary but many other companies are. This is an old pre-conceived view.

Also, I thought that the industry has started to understand that CPU speed i.e. GHz is not a good indicator of performance, with all the claims that IBM is making I would like to say that the equivalent amount of T5220 (8 core, 64 threads with 1.4GHz) servers that take up the same floor space, power and cooling as a IBM z10 will be able to can consolidate more smaller systems while maintaining performance and improving availability and management than a z10. Solaris has in-built virtualisation with LDOMS and containers that is free and open source. Now what would the IBM z/OS costs for this be, people really need to look at this area?

Strangely I just read in this article that IBM has caught up with large UNIX mainframes with their z/OS mainframe. However, this article got some very simple technical details wrong. Large Sun Enterprise Servers (which have all the features of mainframes and more) e.g. the M8000 and M9000 series use up 64 x 2.1GHz and/or 2.4GHz dual core CPU's. You can have up to 64 of these dual core CPU's in one system. Now what we can do in these Sun UNIX mainframes is mix CPU speeds, so if by any chance we have a new faster CPU becoming available in the next 6 months, you can add that to the system while it is running. Now that is innovation and investment protection.

So this claim from the article; "That extra clock speed is important since the 4.4 GHz clocks IBM can now deliver in the z10 give it an edge over the 3 GHz or so clock speeds of X64 alternatives and the 1.6 GHz to 1.8 GHz clock speeds of the Itanium and UltraSparc-IV+ processors used by Hewlett-Packard and Sun Microsystems in their big iron Unix boxes." Is just becoming less relevant.

So IBM is comparing a new z10 system with some old Sun System. These M-series large UNIX mainframes have been available for 10 months now.

As a colleague Jeff Savit wrote in a very good article, IBM is comparing their new products with competitors old products a very sneaky trick and not very honest. Thus this shows that we have to be very cautious about any of the claims in this announcement. This article is a must read for anyone who wants to understand the IBM z10 systems claims in comparison to Unix mainframes.

There is another very strange performance comparison going on here, so that the new z10 can look better, This article implies that the old systems are now specified as being a lot slower than they really were. From the same article here is the quote. "IBM's documents for the z10 say that the fastest z6 core is expected to deliver 62 percent more performance than the fastest z9 core, which had a much lower clock speed. (That would seem to suggest that the 1.7 GHz z9 processor core running full out should have been rated at 568 MIPS, which is lower than the numbers a lot of people have been throwing around for three years.)" So not only does IBM not compare latest z10 mainframe server with old Sun servers, but they compare and reduce the performance numbers of the older mainframes. If I was a IBM z/OS customer I would ask for a refund on the old system software licences as the MIPS value has just changed, since the old z-series was not as fast according to the MIPS indicator as IBM said it was.

Summary, the truth is coming out out slowly, so buyer beware, look at these mainframe claims very, very closely and be suspicious.

I have the greatest respect for IBM, but there are just too many strange claims being made here. After all IBM together with Toshiba (I think) made the cell processor that we have on the Playstation 3. Now that IBM cell chip is really interesting. I cannot say the same about the z10 mainframe with all thes strange comparisons going on.

Just as we expected, the best place for the new mainframe is to replace old mainframes, not Unix (Linux) or other non-mainframe systems. This comes from a end user/customer who admits exactly that, you can read it in this search datacenter article.

But here is the quote "Though IBM has pushed hard to present the mainframe as a consolidation target for hundreds, if not thousands, of virtual Linux servers, Handera saw the new machine as a boon to better handling more traditional mainframe applications such as DB2, WebSphere and CICS batch jobs."

So get rid of the old mainframes and migrate them to the new one. That way you can save costs. During the migration project, you may be able to take some applications off the mainframe and run it on another more open and cost effective and open system giving you more choice and options in the future.

As you move from an old mainframe you often find systems that do not need to run on there anymore, faster, newer and more cost effective Unix servers for which you can do open bids are often more efficient. So it is good to do a review now and again on how you can reduce your zOS dependence.

Remember if you look for a new mainframe, only one company can give you a price. If you want to buy a new mainframe/Unix server you have several suppliers.

Mainframe market is a monopoly, thus you will pay monopoly prices. Do not try and stay dependent on a single supplier, this is not good for business. Shame that we do not have Amdahl and Hitachi making zOS mainframes/servers anymore.

The IBM mainframe is good for running traditional MVS or z/OS apps like CICS and batch jobs that no-one can understand anymore. If you want more mainstream IT stick with UNIX mainframes.

Well we have an old, new mainframe again. I spent a lot of my time with mainframes but moved on when I saw the mainframe market going down, I also felt that I had done as much as I could so I moved to working with Unix, specifically Solaris and HP-UX.

Now there has been a large amount of FUD about this announcement. Some thing that I would like to put straight.

Consolidation - The Mainframe is good to consolidate other mainframes but not really other servers. So if you have 2 or 3 old mainframes that you cannot get rid of. Consolidate them into one of these new ones. Make sure that IBM can show that it will be cheaper to run one mainframe than 2 or 3. Then check the numbers very carefully and do your own research. About consolidating Unix (Linux) servers into a mainframe, now that is a big leap of faith. People should be very careful with this proposition. Should not IBM get more applications developed for z10 and zOS, Open Source z/OS to get better mainframe adoption. This IBM linux consolidation sounds like a desparate cry of help for the mainframe.

IBM mainframe consolidation problem number 1, if the z10 system needs to be serviced including an outage, you will have all your Linux apps out of action. Now IBM will then say buy 2 mainframes, oh dear. Best to stay where you are and consolidate in smaller more manageable and serviceable chunks.

Mainframe big/fast box - I do not believe this for one moment. The large Sun and IBM Unix servers are faster and probably a lot smaller than this old/new iron.

Reliability - The mainframe lost this years ago, Sun and other High end Unix mainframes/servers are just as reliable. The mainframe may be more reliable as you only run your old apps on them which you never change as no-one understands them anymore. But then which server would not be reliable, if you never interfere with it. There are many examples of Sun servers running for several years without a reboot and departments forgeting about them as they "just run forever".

Scaleability - Sun and other large mainframes/Unix servers scale to equivalent or larger systems. Nothing new here for the z10.

Performance - The MIPS algorithm or Meaningless Indication of Performance. Is really a red-herring, you can compare a old mainframe to an new, old mainframe using MIPS but it is not really useful for comparing mainframes with other server platforms. TPC-D type benchmarks are better than MIPS, but still is not a definitive comparison. Companies do not make money running TPC benchmarks they make money running applications, which can only partially be compared using benchmarks.

Innovation - Presently from the information that I have available, the new Z large server (aka mainframe) uses 2 dual core P6 CPU's on a module/circuit board to make it look like a 4 core system. Well you can fool some of the people most of the the time but not all of the people all of the time.

I see that I will need to spend some time looking at this mainframe stuff in detail.

I just worked out why I am having trouble deciding what OS platform I want to go with for the next 3 yrs, after using Linux variants for the last 8 yrs, there is nothing new for the user. I feel that the Linux market has matured with one dominat supplier, Red Hat. With Ubuntu being the only new popular distribution on the market.

So Linux is in a mid life crisis, other distributions did not last this long, my old favourite Mandrake has gone. Suse, I just do not hear about. It is Red Hat this, Red Hat that, all the time. When I ask people what are they using they say Linux, so I ask them, so you make your own and they do not know what that means, 99% of the time people answer Red Hat. Does IBM put together a nice Linux distribution no, they resell others. Does HP create theie own distribution, no they resell others. OK nor does Sun, but I am writing this on a old Sun created distribution of Linux. So the only real creatives in the Linux world are Red Hat and Ubuntu, Red Hat is dominant, smoothering the rest. So Linux is Red Hat in reality.

We should report the Linux market by distribution not by the total, there does not seem much left. Red Hat, Ubuntu or nothing.

This is not too dissimilar in the UNIX server market where it is Solaris, AIX or HP-UX.

Is there some natural law where the industry or market place can only sustain 3 versions of a platform.

In the database market it seems to be Oracle, DB2 and MySQL.

I have a 3yr OS upgrade cycle on my PC, and have been running Sun's Java Desktop System 3 for 3 yrs now. My options are Red Hat, Apple, Ubuntu or I may try Solaris Developer edition, well that will be something new.

Are we searching for eternal youth by trying something new ?

As discussed earlier I am in a quandry about what I should move my laptop to next, when I first started using (Linux) Red Hat around 1999, I did get a wow factor, it was more stable than Win 98, it was faster than Win 98. Linux improved my productivity. Now Win XP is as good as Red Hat, Suse and Madrake seem to have waned a bit and are not as popular.

It just seems that Linux is either Red Hat or nothing, only Ubuntu seems to have any traction or wow factor.

It looks like Apple OS X, which is a Unix BSD variant is the latest wow platform. My wife is an apple fan and long term user so I have to use apple too. It did take me a while to hack into root on an apple, but when I figured all this out I could really fix things, like emptying the wastebasket from files that would not empty. Now and again we need a big hammer to fix things.

Actually it was StarOffice that made Red Hat usable, without an office suite Red Hat, Suse or Mandrake would have never made it onto my laptop. Probably StarOffice (OpenOffice) is what has made Linux successful on desktops/laptops. People say that Staroffice/Openoffice is terrible, however I moved to it to get more stability and I have not used Word any longer, I even do not know what MS office looks like.

Recently there were some articles that Linux will not be sucessful as it is free and big discussion on slashdot

Must be a topical issue, just when I am deciding what to run for the next 3 yrs. A decision not to be taken lightly as I have to spend time fixing my families computers.

I started using Linux as an early adopter back in 1999 when Win 98 just crashed all the time if I was using more than 3-4 windows. As a productivity tool Win98 was a inhibitor. Colleagues would boast about their patching methodologies for Win98, I kept thinking, why should I patch so often, don't they have anything better to do. So a Unix in the name of Linux was appearing for the the desktop and I gave it a try.

My first Linux system was on a Dell Cpi-A Lattitude Laptop, Pentium II 363MHz, 158MB RAM, 4900 rpm 6GB disk, that was with Red Hat.

About 2003 the Dell failed after 2 screens broke and I got a New laptop Toshiba Satellite 2450-S402, Pentium 4 2.5GHz, 512MB RAM,Nvidia GeForce4 420 graphics card, 7200 rpm 60GB disk.

Then I went through the following: Mandrake, Suse, Sun Linux Java Desktop.

Time has come to try something new as I have been running JDS for 3 yrs now.

When I first started it was all about open source, now no-one really makes their own Linux with "make" etc, everyone uses a specific distribution.

Should I use Ubuntu, how different is that, looks OK on my new home PC, but for some reason it reminds me of SUSE, it is the GUI or something.

I have learnt all I can, fixed Nvidia drivers, and Wifi. Power management is not that good on Linux but I do not want my laptop on standby, just in case it overheats in my bag and it starts a fire while I am travelling.

Does not seem like much new is happening in the Linux area, I have learnt all there that I wanted too, the innovations going in are really ports from other places, I understand linux is taking Dtrace and ZFS from Solaris.

What comes after desktop Unix/Linux or precisely, what comes after Red Hat. Looking for inspiration.

A long time ago, laptops were invented so that people could work while out of the office, this was meant to improve productivity. People would upload & download data/email via modems, then broadband or wired connections. Nowadays in the most progressive countries there is free Wifi everywhere. However, the portability of data has caused a problem, security or lack of security. I just read that a UK organisation as documented here http://www.silicon.com/publicsector/0,3800010403,39169759,00.htm has laptops for some of their employees, but will not let them remove these from office, similar problems are ocuring all over the world, military organisations are often losing or misplacing their laptops, data from disks is being lost.

So my logic is such, we have mobile devices which are not meant to be mobile, so why don't we have a more secure and cost effective work/IT model. Historically, organisations went from terminals (keyboard + screen) with no data, desktop (large PC) with distributed data, laptop/notebook with lots of data. Now, we have so much important data on these devices that they are becoming a liability. Due to, too much sensitive data being lost and therefore by implication, possibly could fall into the "wrong" hands or the data could be missused.

As always we have answers or options for these.

1) Encrypted data on all laptops
2) Thin desktop with no data or as we say, stateless devices.

As a Sun employee, we have solutions for both of these.

1) Encrypt data on disk and keep it's integrity, ZFS this is open source and many companies are adopting this outside of UNIX,
2) Companies gave people laptops on their desks, but as they are never allowed to remove them. They are the wrong device, you probably would feel better if the device has to remain at work with a big screen and a larger more ergonomic keyboard. Hence the SunRay.

So all the security people scratching their heads need look no further. Remember the SunRay can run Windows and Unix based applicaitons. Linux to me is Unix, so that includes Linux too.

Picture of the SunRay here: http://www.sun.com/sunray/sunray2

Now I did talk about this a while ago. "A ray or enlightenment in a dark IT quagmire", cost $249 for a Sunray, Suntan lotion not included. http://blogs.sun.com/ValdisFilks/entry/wyse_copies_sunray_idea

I honestly had nothing to do with the article http://www.theregister.co.uk/2008/02/05/ibm_bluegene_web where IBM describes that they can host the whole internet on one "system". As you can guess I do not work for IBM, so I have no great interest in agreeing with them. But I did publish my commodity article way before The Register. However, as a IT person we cannot ignore that IBM does do some good things now and again. Especially when the competitors encourage them to improve. Well that is what market forces do when we do not have monopolies.

Some quotes from the article where IBM agrees with me, or I agree with IBM. You decide.

"So, you're working with systems in a sense very similar to the individual x86 boxes that make up most clusters. Although the unique packaging of the Blue Gene systems along with their low-power cores allows IBM to create a more reliable computer - by more than two orders of magnitude - than commodity boxes which fail all the time."

Now I have only been working in IT for 22 years and seen lots of trends, fashions and designs repeat themselves. A bit like the downsizing, upsizing, rightsizing and funsizing processes we go through. Now am I the only one with the opinion that commodity computers as used in clusters may not be the best solution, if I am then so does the writer of this article. Who said:

"You can't help but get the feeling that IBM and others are on the right track by exploring these hybrid models which place an emphasis on low-power chips and tight, SMP-like design where needed. Maybe we'll all look back at clusters and laugh in a few years."

Could this be the voice of experience.

But seriously, I do like clusters, laptops, PC's, large servers. There is correct place to use all of them and a wrong place. When trends become fashionable the IT business starts to use the latest fashion for the wrong purpose. Then after some spectacular failures we get sensible again. Technology corrections I call them, just like corrections in the financial marketplace.

Well as I work for Sun I can shamelessly plug the fact that if you want to have a systems that you can upgrade while they are running to very large sizes,such as 2TB of RAM 256 cores and 737GB/s memory bandwidth and 244GB/s I/O bandwidth. Then look no further here it is http://www.sun.com/servers/highend/m9000

You also do not have to do the cabling yourself to connect all the CPU's/commodity boxes, it is all done internally by the backplane. This way your datacenter will look less like a spaghetti factory.

Ok I say within 3 years time we will have even larger SMP systems where you could have a present day 4096 cpu cluster in 4 large SMP systems, well that is a statment that puts me in between a rock and a hardplace

Going back to the fact that IT should be exploited as a differentiator to give companies a competitive edge. This is the last of three blogs in my series that computers should not be treated like commodity potatoes.

So take this scenario, we get a load of cheap computers, they fail, someone needs to monitor the alerts, your customers have to have a degredation in service for a while, someone needs to unplug the cables from the failed commodity server, needs to order a new one (which with commodity servers will never be the same as the one replaced), then someone needs to cable it back correctly and install the software. Hopefully install the s/w automatically and hope that the drivers work the same way as the old failed cheap system, as you are using commodity computers the replacement will probably be different than the one replaced. To get to commodity equipment the manufacturers use the cheapest components available at that time, changing the type of equipment used within the computer every month.

There is an environmentaly cost here too. Would it not be better for the earths resources to have one server that lasts 3 years rather than two servers that last 18 months each. You have to add the costs of disposal and the environment to using commodity computers. So after doing all this what have we saved. If the commodity devices are on site in stock waiting to be used then we are wasting money on unused assets and stock, plus the cost of storage, and inventory control.

Think of all the transport costs you are creating in the excess transport of all the commodity equipment that you needlessly replace and move around. Bad CO2 creation, bad CO2, just like bad carbs. Any company that uses commodity IT needs to think about the knock on affects of using disposable commodity computer, if they want to be a responsible organisation. A computer is not just for christmas.

Should we charge companies for the disposal of computers, should companies who like to say that they are environmentaly friendly in their annual reports stop using disposable/commodity equipment. Rather, companies should say and work to a sustainable IT policy, where they are not using lots of commodity items with short lifespans. But use better quality computers with upgrade paths and longer lifespans.

Have we just done some fancy financing and hidden the IT mainteance costs for commodity computers in the christmas party fund. Next time there is no sheery and mince pies at the christmas party, remember the food budget went to pay for commodity computers and you are now eating the cheapest food and drinking the cheapest sherry. Surely not, not in todays sophisticated economy.