ComputerWorld reports that IT professionals are blaming poor training of MCSEs for the recent spread of the Code Red IIS worm, indicating a lack of security training in the program.

– By Jeff Field –
Calling this board the DRAGON is not merely Soyo’s attempt to give a cool name to a motherboard, but is actually an acronym for DDR SDRAM, RAID controller on board, Audio on board, Graphics (representing the AGP Pro slot), Overclocking and Networking — all of the features offered with this board. Read on to find out if this board lives up to its name.The board and expansion
I’m not really a fan of integrated boards — either there’s too much integration, such as with Tyan Thunder boards, or they have very basic features such as AC97 audio, which is one of the first things I turn off. I can’t really recall any integrated board that ever fit all of my needs — until now, that is. The Soyo K7VDRAGON has hit the integration sweet spot. The board features integrated digital audio, IDE-RAID, networking, and overclocking, not to mention a non-essential but stylish design with plenty of visual appeal. Check out the Dragon in action.

PCI expansion on the Dragon doesn’t stand out from the crowd, featuring five PCI slots, the same configuration for newer Athlon-based boards. AGP capabilities are represented by the AGP Pro slot, allowing the motherboard to deliver all the power needed for AGP Pro. While not as fast as AGP, AGP Pro provides more electricity for power-hungry video boards. In addition to the two standard ATA 100 ports supported by the VIA KT266 chipset, a pair of additional ports powered by the Promise PDC20265R IDE/IDE-RAID controller allow up to eight IDE devices to be installed on the board at once. I imagine this should be plenty for most users.

As for external expansion, the Dragon has some standard connectors on the ATX IO panel — two USB ports, two serial ports, one parallel port, PS/2 keyboard and mouse ports, and the connections for the integrated sound. In fact, the only unusual feature on the IO panel is the integrated NIC connector, located just above the USB ports. The board differs from some boards in a couple of other places, however. The Dragon supports a total of six USB ports, including the two on the IO panel, allowing you to make use of front-facing USB connectors popular in new computer systems, as well as add more USB ports in the back. This is a nice feature to have, as USB devices continue to replace older peripherals (this is coming from someone who still types on an IBM Model M). The other unusual external connection is the Digital Audio daughterboard, which has the extra outputs needed by the onboard sound.

Layout and design
The layout of the Dragon isn’t as crowded as it sounds, despite all those integrated components. There are a few problems, however: If you plan to make use of the second IDE controller and full-length PCI cards, you can almost certainly count on being able to use only two of the five available slots; the other three will be blocked by the IDE cables. To test this I installed an Adaptec 3200S (picture), quite a long card, and two IDE cables. The cables had to run to the right, and then to their destination, in order to get there, reducing cable length a couple of inches, but otherwise not impairing function. Not too many people have more than one long PCI card, so using the last two shots should solve this. The other possible problems are the distance of the ATX power connector from the top of the board — large heatsinks could get in the way of the cable.

The next problem, and somewhat more serious, is the placement of a capacitor between the CPU socket and the DDR SDRAM slots. This capacitor can and will block larger heatsinks, but average heatsinks like my Taisol cooler or the Dragon Orb 3 shouldn’t experience any problems. Speaking of heatsinks, my one minor problem with the design is that the heatsink does not have thermal grease between it and the chipset. Soyo includes a packet of thermal grease with this board, so I would suggest that be the first place you put it, overclocker or not; the thermal grease has no downsides and helps keep your chipset cool. Due to the placement of the capacitors, the CPU lever can be difficult to use. Tread with care; damage one capacitor, and you will ruin the board.

As for the appearance of the board, Soyo has clearly decided to push this board to the overclocking, case-modifying type of user — and I have to say, did a good job. With a black PCB and the dark purple PCI slots, Soyo has created a motherboard that is not only completely functional, but as stylish as a motherboard can be.

Documentation and configuration
Soyo documentation has always been slightly lacking, normally a mere 20-page booklet with very basic information on the hardware. The Dragon’s documentation is a completely different story, with 200 pages almost evenly divided between technical information on the board itself and the other peripherals, such is the onboard sound and IDE-RAID. The documentation is both thorough and well written, and covers everything in detail, including the BIOS settings. All of this is a real improvement for Soyo, and I hope this is not merely a special case. All motherboard makers should strive to make documentation that is this good.

Configuration of the board is well covered in the manual. One setting I’d suggest while you are configuring is to set the performance to “maximum.” Doing this increases performance quite a bit and I didn’t encounter any stability problems while using this setting.

Onboard Devices
Promise IDE/IDE-RAID Controller
The Promise controller on the board is labeled PDC20265R. This controller provides either IDE-RAID or normal IDE functions depending on which setting the IDE-RAID jumper is set to. In testing the IDE-RAID under Linux 2.4.7, I found that Linux could not properly address RAID arrays created with it, which is somewhat of a disappointment, but the IDE controller itself functions perfectly under Linux, allowing you for double the number of IDE drives of a normal board.

Cmedia CMI8738 Onboard Audio
Calling the CMI8738 simply “onboard sound” seems unfair; that’s like putting it in the same league with AC97 generic audio. For onboard sound, the Cmedia on this board is nothing short of exceptional. The most exciting feature for me was the Optical Digital output on the Cmedia daughtercard, delivering optical digital out and in and SPDIF out and in, as well center/sub and rear analog outputs. Optical and SPDIF outputs allow for output to home theater systems or devices with digital inputs.

Optical output really excited me because I have a home theater setup, and the optical digital output lets me output from my PC to my HT receiver without the problems of interference I have always had. Optical cables use light to transmit data, so as long as they are not badly bent or kinked, they will deliver a perfect digital signal. The only thing this lacks is Dolby Digital 5.1 output — optical output is limited to stereo output, though it is the best stereo output I have heard from a PC. Perhaps in future Cmedia chipsets, we’ll see Dolby Digital encoding for 5.1 signals from a PC to a home theater, but for now the Cmedia is excellent, it has features you might expect on a much higher-costing board. For Linux support, there is support in the kernel, and Cmedia actively supports its products under Linux.

VIA Rhine 10/100 Onboard Ethernet
Onboard ethernet is something that, like audio, I expect to become commonplace on motherboards. This is mostly due to the proliferation of broadband Internet access methods, such as cable or DSL, that use the standard ethernet connection to provide their connections. The onboard VIA Rhine ethernet was supported by Mandrake 8, and is supported in 2.4.7 as well, functioning fine under both.

Overclocking
While I don’t overclock my systems, I have read a number of success stories of overclocking the Dragon, with a few known problems. First, once you get above 145MHz or so on the system bus, USB on the board stops functioning. Soyo will be addressing this in a BIOS update in the next several weeks. Next is a problem with the CPU voltage adjustment; any voltage adjustments greater than +.100 volts will revert back to the default voltage, another problem being fixed in the BIOS.

The last issue with overclocking is a physical one, and was mentioned above — there is no thermal grease between the heatsink and the chipset on this board, so I would advise anyone, overclocker or otherwise, to use the included packet of thermal grease between the chipset and the heatsink. The actual overclocking features on this board are excellent, allowing for 1MHz increments in front side bus speed, and control of the CPU voltage.

Once Soyo works out these few issues, I would suspect this will be an excellent overclocking board, as of now, it is a very good one — the BIOS update should make it great. The highest I got, with no special cooling, was 1523 MHz, completely stable, and with some minor effort, I suspect it could go a good bit higher.

Performance
System Specifications
Athlon 1.4GHz (266FSB)
256 Megs PC2100 DDR SDRAM from Crucial.com
Western Digital 7200 RPM 10.2 gig hard drive
3Com 3C905TX-C 10/100 NIC (PCI)
400 Watt Future Power ATX 2.03 Power Supply
Gigabyte GV-GF3000 GeForce3

Mandrake 8.0 with Kernel 2.4.3-mdk or custom 2.4.7 (used for sound testing)

Results labeled “maximum” are using the maximum performance setting in the BIOS under Soyo Combo Setup. The result of 1523MHz is a 1400MHz Athlon at 145 x 10.5, default voltage, with maximum performance enabled.

Kernel compiles
In order to test both the board’s stability and speed, I ran three sets of Linux kernel compiles on this board. One is a normal, “uniprocessor” make, or make -j1, which is the default. This uses one process, and does not always maximize system usage. I then did make -j2, which spawns a second process. The last test I run is with make -j3, spawning two extra processes. I do this for several reasons — to find the “sweet spot” for the board/CPU, as well as to stress the system as much as possible when trying to rate its stability. Also, the kernel is extremely useful as a measure of integer performance. The kernel is compiled with all “default” settings.

Kernel 2.4.6 Compile Times (Minutes:Seconds)
Board	-j1	-j2	-j3
Soyo DRAGON	4:35	4:31	4:30
Soyo DRAGON – Maximum	4:28	4:21	4:21
Soyo DRAGON – 1523MHz	4:02	4:03	4:02
Gigabyte 7DXR	4:29	4:20	4:21

Here we see several things. First of all, the “maximum performance” setting does indeed make a large difference — decreasing a kernel compile by as much as 10 seconds. Also, we see just what kind of numbers you stand to gain from overclocking, and this overclock was done with no special cooling — its as easy as changing a setting in the BIOS. Also, once the maximum performance setting is turned on, the board easily holds its own against the 7DXR.

POVRay Benchmarks
POVRay is a multi-platform raytracing program. It is a floating point-intensive task and serves well to help measure the floating point performance of a CPU. For more information on this benchmark, head to the official POVBENCH homepage. The command to run for this benchmark, once you obtain POVRay, you run povray -i skyvase.pov +v1 +ft -x +mb25 +a0.300 +j1.000 +r3 -q9 -w640 -H480 -S1 -E480 -k0.000 -mv2.0 +b1000 from the command prompt. Results are in seconds.

POVRay
Board	Result
Soyo DRAGON	15 seconds
Soyo DRAGON – Maximum	15 seconds
Soyo DRAGON – 1523MHz	14 seconds
Gigabyte 7DXR	17 seconds

Again, we find the Dragon holding its own easily. The maximum settings have little effect on POVray because that setting mostly deals with RAM timings.

Quake 3 Arena Timedemos (Frames Per Second)
Board	640×480	800×600	1024×768	1200×1024	1600×1200
Default Quality
Soyo DRAGON	174.8	174.2	165.9	124.6	90.3
Soyo DRAGON – Maximum	186.5	185.0	172.4	125.9	90.7
Soyo DRAGON – 1523MHz	199.9	198.4	179.9	125.3	90.3
Gigabyte 7DXR	183.6	182.6	172.8	126.1	90.7
Highest Quality
Soyo DRAGON	173.9	172.2	143.6	97.3	71.4
Soyo DRAGON – Maximum	184.1	179.1	145.6	97.9	71.6
Soyo DRAGON – 1523MHz	198.5	190.4	146.2	97.5	71.3
Gigabyte 7DXR	183.2	179.0	146.1	98.0	71.4

We find that with the maximum performance setting the Soyo easily holds its own. We also find what exactly you gain from overclocking. In the case of Quake 3, you gain a lot in the high resolutions, but very little, if anything at all, in the higher resolutions where the graphics card, even a GeForce3, is limited.

bonnie++ results
Bonnie++ is a hard drive benchmark that tests the writing and reading
from both a single large file (such as that of a database) and many small files (like a proxy, or mail program). It is
useful for simulating performance under such applications.

Bonnie++ results
Sequential Output
Controller	Per-Character	Block	Rewrite
Soyo DRAGON	15446 K/sec, 97% CPU	28800 K/sec, 21% CPU	7180 K/sec, 4% CPU
Gigabyte 7DXR	8608 K/sec, 57% CPU	20787 K/sec, 20% CPU	5168 K/sec, 4% CPU
Sequential Input			Random
Controller	Per-Character	Block	Random
Soyo DRAGON	10504 K/sec, 58% CPU	21853 K/sec, 9% CPU	165.7 Seeks/sec, 0% CPU
Gigabyte 7DXR	11187 K/sec, 61% CPU	21949 K/sec, 14% CPU	111.2 Seeks/sec, 0% CPU
Sequential Create
Controller	Create	Read	Delete
Soyo DRAGON	18025 /sec, 96% CPU	none	24256 /sec, 100% CPU
Gigabyte 7DXR	18752 /sec, 99% CPU	none	22857 /sec, 100% CPU
Random Create
Controller	Create	Read	Delete
Soyo DRAGON	19164 /sec, 97% CPU	none	15439/sec, 77% CPU
Gigabyte 7DXR	15559 /sec, 86% CPU	none	19043/sec, 99% CPU

Here we see the Dragon performing well, and neither board seems to be a clear “winner” — both boards do equally well.

HDParm
HDparm tests the maximum data transfer rate of a hard drive in using two methods, uncached (but buffered still by the hard drives on board buffer) and cached (buffered by the drive and cached via the operating system cache). While the uncached test should not vary between different controllers that support up to the drive’s ATA-version (such at ATA-66, which is what this drive uses), the cached test varies between boards because it is essentially a test of the CPU, cache and RAM on a system. This is what makes it interesting in this case — it can help showcase the memory performance of these two boards.

hdparm -t (Uncached) Results
Soyo Dragon	23.70 MB/sec
Gigabyte 7DXR	23.88 MB/sec
hdparm -T (Cached) Results
Soyo Dragon	209.84 MB/sec
Gigabyte 7DXR	220.69 MB/sec

Again, no clear winner here, both are good performers, with the Gigabyte’s theoretical maximum (if all conditions are optimal) being 220.69MB/second, 10 megabytes a second greater than the Dragon’s.

Conclusion
The Soyo Dragon is on the most impressive motherboards I have used to date. Rarely does a motherboard impress me as this one has. Soyo has combined excellent documentation, easy setup, a good design, stability and excellent features into a very affordable motherboard. Not only that, but all the components work wonderfully under Linux, except the IDE-RAID controller in RAID mode, and that is a driver issue.

This is the first truly excellent Socket A motherboard I have seen. If you are planning on buying a Socket A board, I definitely recommend the Dragon; it looks good, its stable, it can be overclocked, and it has plenty of features. The SY-K7VDragon is available for $159 from SpecialtyTech.

For discussion of this review and any other hardware-related topics, please join #Hardware on OpenProjects.net.

Kelly McNeill writes “High-tech companies have spent the last 10 years producing a speculative “bubble” that recently collapsed due to Federal Reserve Chairman Alan Greenspan’s interest-rate errors. This bubble was due to the subtle attempt to get something for nothing — by employers, employees and stockholders. Mr. Greenspan was fooled into raising interest rates in the 1999-2000 time frame by the high-tech industry’s false claims of a “shortage” of high-tech workers. Well, it seems like there is no shortage after all. But Mr. Greenspan was fooled into believing that labor-wage inflation was imminent”

Computer World reports that Tulane University is investing $1.7 million to upgrade its campus network to wireless 802.11b, but is using hardware that can be quickly switched to the 54Mbps 802.11a wireless protcol as soon as it becomes available.

MSNBC reports that NASA has launched a flying wing which has broken the altitude record for non-rocket powered aircraft. The solar-powered propeller-driven flying wing attained a height of 96,500 feet.

Noel Davis at O’Reilly.net warns of “buffer overflows
in Linux telnet daemons, IBM AIX telnet daemons, the Kerberos 5
telnet daemon, Window Maker, and Solaris’ xlock; temporary-file race
conditions in AllCommerce and rcs2log; and vulnerabilities in ZyXEL
Prestige 642R and 642R-I ADSL routers, groff, OpenLDAP, fetchmail,
UnixWare Package Tools, docview, and ColdFusion Server 5.”

Kelly McNeill writes “But what is a company like Palm to do as it faces an increasingly popular and legitimate threat from the PocketPC? What can either company do to be more aggressive in a burgeoning market without being considered anti-competitive? Most PDA users agree Palm has been the superior OS for several years. A few would also admit, though, that having a nearly 90 percent market share — a virtual monopoly on the PDA market — hasn’t exactly spurred innovation on Palm’s part. So here we see an opportunity for Microsoft to show the marketplace how its OS for PDAs better serves the market.”

Slashdot is discussing the news that Loki, a company which ports popular PC games to Linux, has filed for Chapter 11 bankruptcy protection.

Kelly McNeill writes, “As a die-hard Linux user, my first actions when I get a new piece of hardware is to reject Microsoft’s EULA, install Linux, and send the Windows CD back to Microsoft for a refund. Unfortunately, to do that would mean forfeiting a good half of my mobile connectivity, unless I could write a driver for the winmodem myself. Instead of attempting this thankless task without support from HP or anyone else, I instead looked around for another solution. What I found was virtual network computing (VNC), a system that allowed me, when my laptop was connected to my home network, to use a Linux desktop hosted on one of my other machines.” The column is at osOpinion.

Seventeen of the world’s top scientists today
supported Princeton University Professor Edward Felten
and his research team’s challenge to the Digital Millennium
Copyright Act (DMCA) on free speech grounds. Prominent
academics, cryptographers, software programmers, and
scientific conference organizers explained to a federal
court the stifling effects of the DMCA on scientific
research and freedom of expression.

The Electronic Frontier Foundation (EFF) represents the
research team in a lawsuit filed June 6 asking a federal
judge to declare that the scientists have a First Amendment
right to publish their research both at the USENIX
Conference in Washington, DC, and elsewhere. USENIX,
fearful of threats that had been made against the
organizers of the prior conference, has joined the suit,
and plans to webcast the presentation of the paper on
August 15.

In his latest declaration to the court, Professor Felten
explained, “I understand that Defendants advocate an
interpretation of the DMCA that would outlaw analysis of
systems that might be used to control the use of
copyrighted materials…. such an interpretation would
effectively prevent analysis of critical systems, and so
would have a disastrous effect on education, research, and
practice in computer security.”

He further commented, “Not only in computer science, but
also across all scientific fields, skeptical analysis of
technical claims made by others, and the presentation of
detailed evidence to support such analysis, is the heart
of the scientific method. To outlaw such analysis is to
outlaw the scientific method itself.”

The case arose after scientists from Princeton University,
Rice University, and Xerox tried to publish research that
reveals flaws in the recording industry’s control systems
for digital music at an April 2001 conference. The
recording industry claimed that a 1998 law called the
DMCA prohibited the presentation of the research paper.
In a series of e-mails and conference calls to the
researchers, their universities, and conference organizers,
recording industry attorneys intimidated the researchers
into withdrawing their paper from the April conference.
Hours after the paper was withdrawn, representatives of
the Recording Industry Association of America (RIAA)
issued a press release claiming that they had never
intended to prevent scientific speech.

Last month, the RIAA asked the court to dismiss the current
lawsuit after sending a letter to the court stating that
it would allow the conference to go forward, so no case or
controversy exists for the court to decide. The EFF legal
filing today rebuts that claim and reveals the chilling
effect felt throughout the scientific community since
Congress passed the DMCA and the recording industry started
threatening researchers.

In legal declarations supporting the researchers,
scientists worldwide expressed concern about traveling to
the United States, where the FBI has already arrested
and jailed a programmer for allegedly writing software or
conducting research that could help someone use a
copyrighted work in ways disallowed by the publisher.

Niels Ferguson, a prominent Dutch cryptographer who
recently discovered major flaws in a commercial
high-definition video system, told the court, “Despite the
fact that I performed all the work in Amsterdam, I could
face arrest if I visit the US after my research had found
its way into the jurisdiction. My research is silenced
since I cannot talk about my scientific results to my
colleagues and peers, as is now the case since the DMCA
became law in the US. Scientific freedom is not only
threatened under this law, it is demonstrably curtailed.”

The DMCA prohibits providing information that other people
may use to circumvent the technological protection measures
placed on digital files. In this case, the recording
industry developed watermark technologies to help control
how consumers can use digital music.

Professor Felten’s team discovered that it was easy to
circumvent the technologies.

“The recording industry has done untold damage by
their threats to Felten and the other researchers, their
universities, and the conference organizers. The resulting
chilling effect on the broader scientific community
continues unabated,” said Robin Gross, intellectual
property attorney for EFF. “We on the Felten legal team
will work to ensure that industry can no longer use the
DMCA to threaten freedom of speech and scientific
progress.”

EFF’s response to the RIAA’s motion to dismiss,
including the supporting declarations and more background
on the legal challenge:
http://www.eff.org/sc/felten/.

RIAA’s motion to dismiss the case:
http://www.eff.org/sc/felten/20010712_riaa_mtd_memo.html.

RIAA/SDMI April 2001 letter threatening Professor Felten
and his team:
http://www.eff.org/sc/felten/20010409_riaa_sdmi_letter.html.

Professor Felten’s website:
http://www.cs.princeton.edu/sip/sdmi/.

Listen to an audio file of the press conference held when
the case was launched [MP3]:
http://www.eff.org/sc/felten/felten_audio.html.

For more information on the August USENIX Security
conference:
http://www.usenix.org/events/sec01/.

About EFF:

The Electronic Frontier Foundation is the leading civil
liberties organization working to protect rights in the
digital world. Founded in 1990, EFF actively encourages and
challenges industry and government to support free
expression, privacy, and openness in the information
society. EFF is a member-supported organization and
maintains one of the most linked-to websites in the world:
http://www.eff.org/.

About USENIX:

The USENIX Association, an organization representing some
10,000 computer research scientists is dedicated to the
free exchange of scholarly information through its many
conferences and publications. See its website at:
http://www.usenix.org/.