The Great Defrag Shootout VI: 4th Generation Defrag

I have suspended my evaluation of mst Defrag Home Edition 2.0 Beta, while I wait for a response to two lengthy bug reports. It is beta software, after all. Don't bother with version 1.9: it sucks, but not nearly as badly as the Ashampoo software. I found some interesting information in the mst help file, which I share as written.
"Fragmentation occurs naturally on a computer system's hard disk as files are created, deleted, and modified. Eventually, the operating system needs to store parts of a file in non-contiguous data clusters. Although invisible to the user, it can slow down the speed at which data is accessed, because the disk drive must search through different parts of the disk to assemble the various fragments into a single file to present them to the user.
"To resolve this problem, defragmenter applications were designed to defragment or 'defrag' the hard disk and reassemble these file fragments into the same and contiguous data clusters.
"There have been four different generations of defragmenters since the early days of computing.
"The defragmentation tools of the first generation could defrag disks only if they had exclusive access to it. This meant you had to choose between work or defragmentation. For example, the application defrag.exe from MS-DOS 6.2 worked in this manner.
"The second generation of defragmenters could share use of the operating system with other applications. To start the defragmentation process the user or administrator had to start it manually. This could often be inconvenient, especially on business computers in large company environments.
"With the next generation three defragmenters it was no longer necessary that a system administrator be logged on to the system. The real work was done by a service (some kind of background task) which had the relevant rights and which could do its work even if no user was connected to the system. Nevertheless, significant problems remained due to the challenges involved in configuring the application properly. A lot of systems remain fragmented because the user simply forgets to run a defrag. In business networks the rollout and scheduling has proved to be a daunting challenge.
"Finally mst Defrag introduced the first fourth generation defragmenter to employ their proprietary 'proactive' defrag technology, which effectively resolves all of the above issues. To install mst Defrag you need administrative rights, of course, just like with any other defrag software. Once installed, mst Defrag will defragment the disks unobtrusively in the background while a logged on user can continue working undisturbed. But only with such a 4th generation defragmenter are the disks defragmented proactively in the background without user or administrator interaction, insuring optimal performance of the user's hard drive at all times.
"The technology which is used in mst Defrag is, like most other brilliant innovations, actually quite simple. The goal in the development of the proactive technology was to relieve the computer system administrator or user from tedious and time consuming tasks.
"As soon as a file is saved by Microsoft Windows in more than one fragment, mst Defrag monitors this and memorizes the file name internally. As soon as there are adequate free resources to run the task, mst Defrag automatically performs the defragmentation of that file. This action occurs proactively (in real time); that means that it is performed completely independently by the software and requires no user interaction. Any running processes will not be impaired.
"Together with an initial defragmentation, which is also done completely automatically after the installation, the optimal state of files on the disk files is virtually assured. And this is maintained not only after the initial defragmentation but but automatically thereafter using the proactive technology."
The shootout continues with O&O Defrag 8.6 Professional. Stay tuned!

The Great Defrag Shootout: Part I | II | III | IV | V | VI | VII | VIII | IX | X | XI | XII | XIII | XIV | XV | XVI | XVII | XVIII | XIX | XX | XXI | XXII | XXIII | XXIV | XXV | XXVI | XXVII | XXVIII | XXIX| winner | all | why

eNatis developer forced to eat his own dog food

ITWeb Johannesburg, 4 May 2007 - Yesterday was another record day for eNatis, its developer says, with 500 000 transactions completed – 68 000 between 2pm and 4pm alone. [That's the time I was standing in a 2.5 hour queue]
However, police had to be called to some licensing stations, such as the Johannesburg Metro Police's facility in Loveday Street, to deal with increasingly restive crowds.
The Department of Transport and the Presidency are now calling on the public to be patient, rather than denying anything is wrong. [They have been denying it from the outset. How about accepting responsibility?]
Commenting on yesterday's statistics, Tasima CE and eNatis project director Johan Vorster says: “This is still not good enough, we need to increase capacity.” [Nice of them to say so, but when do they apologise on their web site?]
He says technicians have overnight started installing a fourth database server to speed up transactions generally and allow eNatis to cope with demand spikes.
The system began operating on 13 April with two database servers. [Actually it was one, and they added a second on 19th April] Demand caused Tasima to add a third server last week and now the fourth should be online by Monday. “This fourth database server roughly has the same capacity as the other three together,” Vorster says.
“eNatis is handling volumes well, but is getting occasional spikes. We realised that on the database side we were marginal, and sometimes, when large requests come through, we get overload, with 100% CPU utilisation on one of the servers. They should be load-balancing, but it seems they get too busy that they don't share, causing the system to slow down.” The fourth server should help in that regard, he adds.
The Tasima consortium, lead by arivia.kom, is also looking at the system's transaction flow and has “appointed data consultants to see where it can be optimised”. [That's about 3 years too late, but better late than never. Any system that needs to shut down for 1 week to make a changeover and convert data is clearly not properly planned or implemented]
Vorster adds he, too, has had a run-in with the system while attempting to renew a vehicle licence.
The Retail Motor Industry Organisation (RMI), which represents about 7 800 dealers, wrote a scathing letter to transport minister Jeff Radebe about the problems it encountered with eNatis.
In the letter, dated last week, RMI CE Jeff Osborne told the minister transactions “of vehicles and equipment totalling more than R1.5 billion cannot be completed, as no vehicle licensing documents can be obtained for submission to financial institutions… Many businesses are at a standstill and unable to trade."
Cabinet spokesman Themba Maseko yesterday told a post-Cabinet media briefing that the country's political leadership has “noted the frustration experienced and calls on the public to exercise patience during this period as the new system will bring great benefits, both to the public and the authorities in the long term".

ITWeb stories: eNatis ‘no bed of roses' | eNatis pulls through | eNatis beefed up | Govt, developer defend eNatis | Technical hiccups hit eNatis | Govt upgrades Natis | Original Announcement 28 July 2003

The Doctor is in ... business

Click to view the detailed image.

The Great Defrag Shootout V: Auslogics Registry Defrag 4.0

While I was investigating the Auslogics Disk Defrag program I stumbled apon a companion freeware product: Auslogics Registry Defrag, currently version 4. It has turned out to be quite useful, and allowed me to uninstall Uniblue Registry Booster once and for all.
It's a picky program, insisting that you stop running everything else, and providing a greyed-out background to your entire screen once it starts. You are forced to reboot when done, although you can cancel the process. I didn't mind these demands because it seems to get the job done. I ignored the request to exit NOD32, my antivirus program, since it doesn't make changes to the registry anyway.
I may have removed the companion Disk Defrag program, but the Registry Defrag program stays put. Now I begin testing mst Defrag Home Edition 1.9 and their 2.0 beta. Stay tuned!
Update: This Registry Defrag program is a beta product, and occupies 4.97MB of hard drive space. I have found another one called NTREGOPT that uses a fraction of the space. Together with PageDefrag v.32 from Sysinternals, I can keep my registry in tip-top shape with tried and tested software.

The Great Defrag Shootout: Part I | II | III | IV | V | VI | VII | VIII | IX | X | XI | XII | XIII | XIV | XV | XVI | XVII | XVIII | XIX | XX | XXI | XXII | XXIII | XXIV | XXV | XXVI | XXVII | XXVIII | XXIX| winner | all | why

The Great Defrag Shootout IV: Auslogics Disk Defrag 1.1.2

The best thing about this program is that it's free. It also has a nifty user interface and appears to work well. That's where the illusion ends. Even Windows Defragmenter works better than this, although probably with less panache.
First, the good points: it has a clean, helpful, intuitive user interface, and does a good job of telling you what files it is busy with, something that PerfectDisk doesn't do. As it processes the files it shows their characteristics on the display map. From time to time it does weird display things, so after a while I lost confidence in what it was claiming to display.
The next good point is the detailed report it displays on demand once it is finished. The information is well presented and could be useful. It is certainly well presented.
Unfortunately the weakest point in the entire program is the defragmentation algorithm itself. It's not as good as the built-in Windows Defragmenter. Considering that Auslogics uses this freebie program to tout the virtues of its PC performance optimisation package, it only succeeded in making me extremely sceptical of the products they sell.
I'll concede that the test it failed was a fairly heavyweight test: I copied a 4GB file from the server into a compressed directory on the local hard drive. It's an archive of a 4GB SQL database, and contained several thousand fragments by the time it was copied. Then I restored this data to the main SQL Server database, again stored on the hard drive in compressed form. Once this was complete, I asked the Auslogics program to defragment the now fairly cluttered hard drive.
I was somewhat annoyed when the package completely skipped this first file, claiming not enough free space to defragment the file, and then crunched its way through the remaining 2 files and a number of smaller ones. The end result was a bit of a mess, with the free space fragmented into a number of small sections. The defragmentation algorithm clearly does not try to consolidate free space, and only looks at existing files to see if they are fragmented or not.
I have uninstalled this package from my hard drive. It may be free, but it lacks some features that even the built in Windows Defragmenter provides, such as a command line interface and the ability to schedule defragmentation automatically. Perhaps version 3 or 4 will do a better job, but for now its not worth downloading.
Update: Version 1.2 will add command line support and scheduling. No idea when it will be ready.

The Great Defrag Shootout: Part I | II | III | IV | V | VI | VII | VIII | IX | X | XI | XII | XIII | XIV | XV | XVI | XVII | XVIII | XIX | XX | XXI | XXII | XXIII | XXIV | XXV | XXVI | XXVII | XXVIII | XXIX| winner | all | why

National Productivity Week

What costs R400 million, is three years late, and makes no damn difference? eNaTIS, that's what. It's just as slow (if not slower) and just as broken as the old system, and it is a convenient excuse to waste everyone's time. Today it wasted 3 hours of my time at the Vehicle Licensing Department in Randburg.
I arrived at 12:05pm, and the queue was out of the building door. I stopped a man leaving the building with a registration paper in his hand, and he told me he'd arrived at 9am, so I hunkered down for a 3 hour wait. By 12:30 the queue had reached the doors and I was able to hand in the change of ownership forms. They were stamped, and I was given number 37 and told to wait for the number to be called. I paid a traffic fine at another counter, and then joined the license renewal queue at 12:37pm. We all stood and waited until 1:40pm, when the game of musical chairs began. They provide 20 seats and you move from one to the next as the queue advances. It took nearly an hour for this queue to advance, partly because only 3 of the 4 windows earmarked for renewals were actually being used.
Once I had handed my form to the operator and waited a few minutes while they backdated the payment date, I was able to pay for the renewal and get the certificate after 2 hours and 30 minutes. So much for 12 transactions per second.
But wait, there's more! I still had the matter of the change of ownership papers, which had been languishing in the back of the office as batch number 37. Finally the number was called and I left at 2.50pm. Why does it take 2 hours and 13 minutes to change the ownership of a vehicle? The mind boggles.
I guess I shouldn't be surprised. After all, this is what I wryly refer to as National Productivity Week, being the week during which we have 2 public holidays optimally positioned 4 days apart to create the maximum disruption to the economy and waste everybody's time.

Naional Productivity Week | Bumpy Start for new eNaTIS Licensing System | New Electronic Licensing System gets Help

The Great Defrag Shootout III: Golden Bow Vopt 8.18

The first time I heard about Vopt was listening to the Security Now podcast. Considering that "Mr Hard Drive" Steve Gibson, of Spinrite fame, was raving about it, I decided to try it out.
First impressions count, and my first impression of their web site was that it was designed 10 years ago, and not much had changed since. Obviously the webmaster is a man of few words. This isn't a bad thing, and at least I was spared the endless chatter and hype of "affiliate"-oriented web sites. No adware, no spyware, no bloatware.
Installing the program was pretty simple, and the initial interface reminded me a lot of the old DOS version of Norton Speed Disk. It loads fast, and shows an initial map of the hard drive without any waiting. First good point.
Second good point is that it runs fast, and the screen display tells you what is going on. I like the way it starts at the beginning and fills up all those little gaps with files from the end of the disk.
Third good point is that the Vopt algorithm tends to maximize the free disk space as it moves the files around. The first "bad" point is that Vopt has a totally different approach to PerfectDisk when it comes to what to do with the Master File Table (MFT). PerfectDisk (PD) and Diskeeper (DK) reserve trailing space after the MFT so that it can grow as required by the operating system. Vopt uses this space for files, presumably on the assumption that the MFT can grow in a fragmented way jut like any other file, and then it will sort this out on the next defrag pass.
The PD/DK approach would make a lot more sense if PD didn't insist on placing the first part of a large file in front of the MFT and the rest of the file after it, deliberately fragmenting the file, albeit in only one place. PD does the same nonsense with other fixed files. Vopt takes a more intelligent approach.
Vopt also warns you about cookies and other junk on the hard drive, and has a most useful little analysis feature that shows you what file is using up a given cluster, as well as a list of all the "large" files, "fixed" files, "fragmented" files, and so on. It also shows the name of the file that it is currently working on, which is most interesting, and a feature sadly lacking from PD.
Vopt has no facility to defrag the drive during boot time**, nor does it seem to give any priority to files that are loaded during boot time, so over an extended period of time the Windows startup process may end up taking longer. Another problem is that it doesn't work too well when eMule is running, since both programs slow down when Vopt tries to defrag files that eMule is using.
It may not be as thorough as PD, but it is much faster and doesn't overwork the drive, keeping the file movements to a minimum. The more I use it, the more I like it. If you've already spend $39 on PD it may not be worth spending another $40 on Vopt, but if you haven't got a decent defrag utility already, and you want the defrag program to work quickly and do a good job, Vopt is a good choice.
If you already have PD and get impatient with its (comparatively) sluggish behaviour, then it may be worth the extra cash. I can't bring myself to uninstall Vopt or PD, because I know that when my drive needs to be properly reorganised every few months, PD is the only one up to the task. But at the same time, Vopt has been keeping my drive in good shape on a daily basis, so I can't uninstall it either.
For the purposes of the review process, I will try the freeware Auslogics Disk Defrag 1.1.2 next. It's going to be tough not using Vopt.
**Update: According to the Vopt technical support people, " it's not true that Vopt8 can't run at boot time. You can use the Schedule Tool to tell it to run at Startup, and it's so fast it will be done if you turn on your computer and go get a cup of coffee."
Update: My attempts to defrag my drive's Master File Table (MFT) using Vopt failed. According to the help file's notes section, "Metafiles, including the MFT (Master File Table), are automatically defragged as part of normal defragmentation. In rare cases, the drive configuration may not permit complete defragmentation of the MFT."

The Great Defrag Shootout: Part I | II | III | IV | V | VI | VII | VIII | IX | X | XI | XII | XIII | XIV | XV | XVI | XVII | XVIII | XIX | XX | XXI | XXII | XXIII | XXIV | XXV | XXVI | XXVII | XXVIII | XXIX| winner | all | why