Archive for category Hardware

inside the Phone Operation Center (POC)

Since we are kind of interested in providing a large user DECT network on the campus of our university we spoke with the Eventphone guys about their setup: and as far as we can tell it’s far more sophisticated than expected.

little panoramic view of the POC server room

the Alcatel machine rear view

patch panel

front view

front view with logo

switching switch

audio patch panel

the actual POC service area in the main hall

Well sorry that there are more pictures than text – it’s just because we have to compile all the information we got and pile it all together into a nice and well understandable text for your reading pleasure.

No Comments

Sanyo XF-45 … 10.000 ANSI Lumen photon cannon.

Hrhr – MORE POWER! Today a SANYO XF-45 arrived for our testings tomorrow…

Well just remember: some days ago Samsung announced a LED projector with humongous 23 ANSI Lumen… compare that to the 10.000 of the XF-45…

Source 1: Sanyo XF-45
Source 2: Samsung SP-P300ME

1 Comment

4500mAh battery for the HTC Himalaya / MDA2

Mhhh… I NEED to get one of those. But $140 is quite a bit if you can get a 1200mAh battery for less than $20.

“With normal usage it was 70.5 hours real time duration before I reached the 10% minimum level at which I needed to put it back on charge, thatÔÇÖs a predicted total of over 75 hours duration and a Pocket Plus measured usage of almost 19 hours – Wow!”


No Comments

vandalism – damn this isn’t france, is it?

Bad things are taking place in Ilmenau these days. First there was the WiFi link of FeM which was stolen whilst we’re at the christmas party….

there was a SOEKRIS PC.

The thieves left the enclosure open so there’s probably water in the cableing. ­čÖü

And as if this wasn’t enough on this saturday some bad guys destroyed two of the server room air conditions of FeM:

server room: I

server room: C

temperature peak after the air condition was gone

The I server room air condition was fixed with parts of the C server room air condition – but the C server room remains with temperatures around 30 degrees after we shut down the not necessarily needed servers. Damn my mail server is down since saturday since it’s a 4-way xeon machine consuming way to much power and producing way to much hot air for an unconditioned server room.

What’s going on people? Hopefully the thieves and vandals will get caught.


so these next-gen consoles are all about circles? Sony vs. Microsoft

We all know the Xbox 360 Ring-of-Light green circles logo below. It’s just everywhere these days. So concentrate on the Ring-of-Light and scroll down:

Xbox 360 Ring-of-Light

So since this weekend the CES 2006 website of Sony is online. At the CES 2006 Sony is supposed to show details and in-depth information about the upcoming Playstation 3. And well the site looks more like if they have to tell us something about the Xbox 360:


This is quite funny – I wouldn’t bet it’s not a marketing gag but if it’s not it’s funny and stupid. Please Sony! More information on your vaporware!


No Comments

some DivX news …

Great news these days: DivX 6 is finally available for the Mac – which is great since the last version was 5.2 available since August 2004.

But that’s not all they have for the starving encoding community (yes, it’s you Karsten!). DivX Labs released a new version of the DivX codec for Windows. And this time they tuned the hell out of it: 14 to 80 percent faster by their benchmarks:

“According to our benchmarks, even a modest single CPU will run anywhere from 14 to 80% faster than the 6.0 codec depending on the quality mode selected. At the same time, brand new support for HT, SMP, Dual Core, and Dual Core + HT CPUs allows for gains of up to 300%!”

Finally multiprocessor concepts have found their way into media codecs – finally because this is what we’ve been waiting for years.

Source 1:
Source 2:

No Comments

listen up Tino Seeber! two new Mac mini and 19″ Samsung TFT displays…

Tino Seeber! Look what we bought with your 40 euro membership-fee! Two new 1.5 Ghz Mac minis with 1 GB RAM and 2 new SAMSUNG 940T TFT 19″ displays.

The only things that where not delivered in time are the wireless mice. (no, NO MIGHTY MOUSE!)

the famous “Mount mac-mini”

reorganising the desktop

ready steady go…

installing Microsoft Office for Mac 2004

For those who don’t know who Tino Seeber is or why I am adressing him in this article: Tino Seeber is a guy who thinks that the only purpose of FeM is to give him cheap broadband access to the internet – that’s why he only wants to pay for his internet uplink which is ridiculous since FeM is a research organisation and not an ISP.


No Comments

bending pcmcia cards…

uhh well cosrahn had the great idea to insert more than one card into his tight PCMCIA slots… you can see… a TV card and a DLINK WLAN card beneath it…

bend it baby!

No Comments

secret intelligence information arrived

Plasmon Raidtech sent a quite interesting marketing brochure. Designed like a intelligence report … photos…some text… whuu.. it even disappeared today… ­čśë

No Comments

storage space sizes for removable media growing…

Finally there are definitive words about holographic storage medias. Maxell made a press release today of it’s upcoming 300 GByte holographic medias – yeah!

“With uncompressed storage capacities achieving 1.6 TeraBytes per disk and data rates as high as 120 MBPs.”

“One 5┬╝ inch-diameter optical disc can store up to 150 million pages – more than 63 times the capacity of DVD. Also, with holographic recording, a multiple of form factors, such as discs, cards, etc., and laser wavelengths (red, green, and blue) can be used. “

“The first generation of holographic media with 300 GB of storage capacity and a 20 MBPs Transfer Rate is scheduled for release in late 2006.”

Source: Maxell-USA

No Comments

Windows Media Center Edition for in-car use

So this is what namenlos is working on at 3soft!?

“The 7″ touchscreen is connected to a computer under the temperature controls. The computer features 512MB of RAM, an 80GB hard drive, wireless 802.11g networking and a slot-loading, dual-layer DVD Burner. If you’re like me, you’re already making a list of the items you’ll need to complete this in your car. Absolutely amazing.”


No Comments

FeM condoms – be smart, be safe, your future depends on it.

Hrhr…. the official FeM personal firewall.

remember…December 1st is world AIDS day….


NetApp test aborted.

Disappointing but true: due to the nearly complete lack of documentation and time we aborted the test of the NetApp FAS270 and DS14MK2 today.

The management software is… well – not self explaining at all. Since we do not have a documentation that goes further than setting up the basics we where unable to create a volume big enough to test with – we even where unable to create a iSCSI LUN/volume at all.

Of course we tried and looked – it’s not as if we’re totally unexperienced with such devices. But all hope that we possibly would see a concept that would self explain what to do to get to the goal was not fulfilled.

Very disappointing – unfortunatly the time we had for testing is over – so good bye NetApp – hello IBM. The next thing to test is the IBM storage solution – hopefully better documented and manageable.

The last words of this article by the NetApp:

fem-iscsi-test> halt
Thu Jan 1 01:23:47 GMT [fem-iscsi-test: kern.shutdown:notice]: System shut down because : “halt”.
Thu Jan 1 01:23:48 GMT [fem-iscsi-test: iscsi.service.shutdown:info]: iSCSI service shutdown

CFE version 1.2.0 based on Broadcom CFE: 1.0.35
Copyright (C) 2000,2001,2002,2003 Broadcom Corporation.
Portions Copyright (C) 2002,2003 Network Appliance Corporation.

CPU type 0x1040102: 650MHz
Total memory: 0x40000000 bytes (1024MB)


we need your help: be a video angel for the 22nd Chaos Communication Congress

This year more than any year before the 22nd Chaos Communication Congress needs your help!

Since FeM e.V. plans to help the team of the 22c3 with equipment and manpower there never can be enough hands to help at such a big event.

If you want to assist and help us recording and streaming the whole congress than surf to the official 22c3 wiki and state your intention to help by contacting the orga-team telling them that you want to be a videoangel.

Source 1:
Source 2:

No Comments

it’s alive!!! the NetApp FAS270 and DS14MK2 is up-n-running… YEAH!

NetApp Release 7.0.2: Sat Oct  8 01:25:43 PDT 2005
System ID: xxxxxxxxxxxxxx(); partner ID: xxxxxxxxxxxxx()
System Serial Number: xxxxxxxxx ()
System Rev: F0
slot 0: System Board
Processors: 1
Processor revision: B2
Processor type: 1250
Memory Size: 1022 MB
slot 0: FC Host Adapter 0b
28 Disks: 3808.0GB
1 shelf with LRC, 1 shelf with EFH
slot 0: Fibre Channel Target Host Adapter 0c
slot 0: Dual SB1250-Gigabit Ethernet Controller
e0a MAC Address: 00:a0:98:02:0e:1a (auto-1000t-fd-up)
e0b MAC Address: 00:a0:98:02:0e:1b (auto-unknown-cfg_down)
slot 0: NetApp ATA/IDE Adapter 0a (0x00000000000001f0)
0a.0 245MB

No Comments

graphic card on it’s way to fail…

Thankfully enough my graphic card seems to announce that it’s going to die in the near future… sad but true: the Geforce FX 5200 passive cooled graphic card is obviously not cooled well enough…

No Comments

HDR what?

“Click here to search a HDR capable cable on”
hmmkay… what kind of wire would that be?

No Comments

Burning a CCD

Do you see the slight pink line in the right part of the picture… that’s what the sun does to the CCD when you are not careful… in this case a NIKON Coolpix 990 CCD sensor just died…


No Comments

thanks for the Wafer…

I held a 3 hour lecture at the “.NET Chaostage” event at the FH Deggendorf last week and I got this really cool present from Helena:

It’s a real silicon wafer from Intel (of course a not usable one with contamination). But cool, eh?

No Comments

Ford Galaxy…killed in action

Last friday I was on the streets as usual on my way to a lecture at the FH Deggendorf (.NET Chaostage, you remember?)…It was a normal ride until suddenly the turbocharger decided to stop working for me…Since I was driving at about 180 km/h at this very moment it felt like getting massive headwind.

killed in action.

So…I called the SIXT service and got a new car…a Ford Focus this time…small very very small. But nice to drive. Infact the Ford Galaxy was way more comfortable …

a turbocharger…

Source: Wikipedia Picture

No Comments

22C3: 3G Investigations – scanning your 2.5G/3G ip network for fun and profit

Ahzf and myself are going to present a lecture at the upcoming 22th Chaos Communication Congress (22C3) in Berlin at the end of this year (27.-30. December). Today the Final notification of acceptance of our proposal arrived…

At the moment I am in the Implementation phase of the project…:

IPv6 on Windows Mobile 5…cool eh?

The final papers need to be ready due to december the 1st…


No Comments

inspected with pride

Namenlos reports that there was a delivery today at the office containing the brand new notebook bag…and inside there was a small note that made us proud…

Source: namenlos’ blog

1 Comment


Florian is working on a XEN setup for our new web hosting machine. It’s a Dell PowerEdge 1850 machine with enough RAM for the four instances…more on that when it’s running productive…


No Comments

networked doorbell many wires do you need to connect a doorbell? You need 8 wires of course! Do we actually have any cables !=CAT5/5e/6 ?


JASJAR cluster…

hrm….Me is happy.

No Comments

OH MY GOD! the Easyraid X8P died?

Well… first there was a 1.3 TByte RAID5 volume that showed several defect sectors…and then it disappeared completely.

Now what? – We changed the internal cabling of the host system … nothing changed. The SCSI Interface Card does see an Easyraid X8P device but no volume shows up in the Disk Management.

Fortunatly I made a copy of all the data to our – now working – Promise VTrak m500i…

No Comments

the Network Appliance FAS270 and DS14MK2 arrived…hurray

28x 146 GB hardrives (Seagate Cheetah ST3146807FC and Hitachi Ultrastar HUS103014FLF210)…I am waiting for my login to be activated at the moment – so I could not do any tests. NetApp decided to not deliver any CD or manual inside the two packages…weird.

The hardware itself looks very very promising (got the joke?)… it’s all redundant and quite heavy. You can clearly see that it’s far more pricy than the cheap promise/easyraid hardware.

Eggs! Eggs!? Eggs?

on top the FS270 and below it the DS14MK2. All equiped with 10k UPM SCSI drives.

As soon as I got all my logins I will post several tests and information about the two devices. So be sure to check back often to

1 Comment

unerring dart game

You can have a lot of fun building cool things when you’re student… and in that particular case they built a dart game that is really cool ­čśë

The dish follows the dart…


1 Comment

doing research is fun…toys-for-boys pt. 2

awww look the two got a little MDA2…aren’t they cute?

1 Comment

VGA resolution is fun on a Windows Mobile device…

Since the JASJAR / MDA pro has a gorgeous 640×480 display there are some tools – originally designed for Windows Mobile 2003 SE to allow you to use the high dpi display with all it’s capabilities.

The normal case is that Windows Mobile scales everything up when it detects a higher resolution than 320×240 (QVGA). Since that means that all text on the screen looks really good but takes the same space like it would on a QVGA.

And there comes ozVGA. It’s a small tool that helps you to configure the systems dpi settings to better match the display. That does not necessarily lead to a better readable display but you’ll have much more space on the display since it’s used in it’s “unfiltered unoptimized raw dpi”-beauty.

before ozVGA

after ozVGA.

There are some bugs to this moment: the status bar at the top is not displayed correctly. Some of the applications cannot cope the high dpi (such as the calendar…) – But I hope that future updates of ozVGA will solve that.


No Comments

it’s JASJAR / MDA pro time… yeeeehaa!

After several weeks of waiting today the JASJAR / MDA pro arrived together with the T-Mobile UMTS contract + USIM.

As you can imagine the unpackaging was quite an event (I am not posting the pictures that show me retardedly smiling)

Everything in place….


Since ahzfs JASJAR also arrived today and the third one is on it’s way and expected to arrive on monady we now can just start over developing the applications for our research project.

1 Comment

What now about the Promise VTrak m500i? Final news.

So I was asked why I would accept another test, after all the unsuccessful tests and support calls. The point is that I wanted to give Promise a second chance because they contacted us on their own talking about concerns they have that we do have some problems with their Promise VTrak m500i.

After all we got our hands on a disclaimer-less firmware update which solved at least our 3 Terabyte problem: We could not create a 3 Terabyte RAID volume. It just did not sync. 2 Terabyte was the maximum volume size.

This bug is fixed. The other ones not. Like the complete crash of the m500i when you try to enable DHCP on the management network interface…

Promise M-Class | Logical Drive 0 Info and Settings

LdId : 0 ArrayId : 0
OperationalStatus : OK NumOfUsedPD : 14
NumOfAxles : 1 RAID Level : RAID5
Stripe : 64KB Sector : 512Bytes
Capacity : 2.94TB PhysicalCapacity : 3.17TB
SYNCed : Yes WWN : 22a6-0001-5501-c7bf
SerialNo : 114163942163624924927000000000000

Alias : raid_3TB
ReadPolicy : ReadAhead
WritePolicy : WriteBack

Hurray! 2.94 Terabyte!!

Promise M-Class | Controller 1 Basic Information

Alias :
Vendor : VTRAK-MB
Model : VTRAK
OperationalStatus : OK
PowerOnTime : 3 hours

CacheUsagePercent : 0% DirtyCachePercent : 0%
CmdProtocol : SCSI-2 WWN : 2000-0001-5500-2027
PartNumber : P29000020000036 S/N : M06G05715100016
Revision : 3 ManufactureDate : Jul 23, 2005
BootLoaderVersion : 1.05.0000.00 BootLoaderBldDate : Jun 17, 2005
FirmwareVersion : 2.00.0000.00 FirmwareBuildDate : Oct 3, 2005
SoftwareVersion : 2.00.0000.00 SoftwareBuildDate : Oct 3, 2005

We actually discovered some software limitations…for example the H2bench which I used for the benchmarks:

H2bench — by Harald B├Âgeholz & Lars Bremer / c’t Magazin f├╝r Computertechnik
Version 3.6/Win32, Copyright (C) 2002 Heise Zeitschriften Verlag GmbH & Co. KG
Dutch translation by F&L Technical Publications B.V.
*** drive too large.

And interestingly there are some limitations in Windows. After I connected the iSCSI device and initialised the drive:

I had two partitions…automatically created. Because the maximum size of partitions on a MBR-Volume is 2 Terabyte.:

“Because partition tables on master boot record (MBR) disks only support partition sizes up to 2 terabytes, you must use dynamic volumes to create NTFS volumes over 2 terabytes. Windows XP Professional manages dynamic volumes in a special database instead of in the partition table, so dynamic volumes are not subject to the 2-terabyte physical limit imposed by the partition table. Therefore, dynamic NTFS volumes can be as large as the maximum volume size supported by NTFS. Itanium-based computers that use GUID partition table (GPT) disks also support NTFS volumes larger than 2 terabytes.”

So I just had to convert it:

and there it is:

   3086914 MB Speicherplatz auf dem Datentr├Ąger insgesamt.
307200 KB in 2 Dateien
4 KB in 9 Indizes
0 KB in fehlerhaften Sektoren
162455 KB vom System benutzt
65536 KB von der Protokolldatei belegt
3086456 MB auf dem Datentr├Ąger verf├╝gbar.

4096 Bytes in jeder Zuordnungseinheit
790250239 Zuordnungseinheiten auf dem Datentr├Ąger insgesamt
790132825 Zuordnungseinheiten auf dem Datentr├Ąger verf├╝gbar

Since there are some problems with the machine left and I am testing it right now you can expect some more news on that subject…

Source 1: Promise Disclaimer Trouble
Source 2: Size Limits in NTFS/Windows

1 Comment

Blinkenlights reloaded

Surprisingly Blinkenlights is back for our viewing and gambling pleasure. If you are in Berlin these days, make sure you’ll take a look.

(it’s the building in the upper middle…which is used as a big b/w screen)

Source 1:
Source 2:

1 Comment

VTrak M500i shutdown -a shutdown

You may ask why we are still testing the VTrak M500i… well… this and more details when I am awake in some hours, until then:

administrator@cli> shutdown -a shutdown

**** Shut down in progress ****

administrator@cli> Shutdown complete. It is now safe to power off the subsystem.

No Comments

a windows iSCSI target software: Stringbeansoftware WinTarget 2.0

We recently tested Stringbeangsoftwares WinTarget Version 2.0. It’s a software implementation of an iSCSI Target with a huge feature set.

“WinTarget Standard Edition provides the core SAN functionality required to centralize, consolidate, and manage storage centrally. The Standard Edition provides this core functionality on an economical basis for those customers that do not need, or are not ready to implement more advanced data management and protection solutions.”

We tested the Standard Edition Evaluation Version 2.0 since we do not need the other features at the moment (like DataView, Snapshot,…).

As usual first the testing setup:

We got 6x 250 GByte Harddisks on 2 Promise FastTrack 150TX4 Controllers. All this on an Intel Serverboard with an 3 Ghz P4 and 2 Gb of RAM. As the client we used the same machine as in all other tests.

14 days should be enough

When you installed it and started it the first time you’ll have something like that:

The Main-Screen of WinTarget 2.0 is divided into the information pane on the right and the section tree on the left side of the window.

WinTarget divides it’s structure into Hosts and Devices. The Snapshot Feature was not tested/used.

We started by creating a new Host:


entering the host name and description

we have to specify which Client IP can access this host…


After creating the host we can create the iSCSI Target Volumes (called “Disks” in WinTarget 2.0):

With WinTarget you can use an actually existing volume or a file to create a iSCSI Target.

We choose the RAID…


a descriptive name…

to which host does the belong to?


Attention: All Data will be deleted from the volume…(we even didn’t have to format it in the beginning…)

there it is…

and it looks good ­čśë

There’s a snapshot feature in WinTarget that we just didn’t test due to limited testing time.

Now everything is setup. That’s it. ­čśë Just that few steps. We can now configure the iSCSI Client:

by entering the IP and Port…

selecting the exported iSCSI Target Volume…

and there it is on the client machine.


in the WinTarget environment it does look a bit different afterwards…

So. That’s it for the configuration. Since we had limited time we could not make a full benchmark of the device. It just took too long since it’s not as fast as expected.

We got around 30 to 40 Megabyte per second reading speed and not more than 7 Megabyte for writing. That’s simply horrible.

We therefore cannot recommand this software combination for our hardware configuration – maybe it’ll be faster with another hardware but as you know: we don’t have the time to test that.



we get full refund: the Promise VTrak m500i adventure is over… don’t buy it!

Today is a great day! We just finished the packaging of the Promise VTrak m500i and it’s hard disks. And soon it’s going to get back to Promise.

As you may have noticed we had trouble with our newly bought m500i. Very serious troubles that all pointed to one source: Promise. The firmware of the m500i just isn’t capable of doing the things Promise is promising. No 3 Tbyte Volumes. No DHCP for the Management port without a complete crash of the m500i. No saved settings (the m500i lost it’s settings after a power outtage). The list goes on.

mount(ed) harddisks:

Now back to business. We await the arrival of a NetApp and an IBM storage solution. So more on that subject soon.

No Comments

protect your property: build yourself a sentry turret

A quite nice job with some very interesting features. Please notices that this is NOT a real gun – it’s an airsoft replica of a FN P90.
These guys made quite an impressive job: Their turret is completly computer controlled and finds and follows it’s targets autonomously.

“Notice the tactical LED flashlight is glowing on the left side of the turret. There is absolutely no practical reason for that.”

Beside the complete building description there is also a short movie of the Sentry Gun in action on the website:

Source: Sentry Gun

1 Comment

BUBBLES! BUBBLES! – the hitcounter YOU want to have.

You have a website and you feel depressed because nobody comments your stupid articles? You want at least a little bit of appreciation? The thing that could brighten up you’re world is: Netbubbles.

This little gadget “monitors” your website: Everytime a user hits your page, everytime someone “buys” or does anything valuable on your website the Netbubble will fire some it’s special-effect-bubbles or even some light effects.

It’s a visual feedback for “the webmaster that cares” (sorry Falko, I had to take it).

dear Netbubbles Company: I WANT ONE! ­čśë I could test it… send me one!!!!1 (I don’t get enough appreciation (e.g. number of comments!))



what’s this raid thing anyway?

You always wanted to know what I am talking about the last days? You got no idea what a RAID is? JBOD does not ring a bell? You think XORing is something sexually offensive? (Hey! You’re right!)

So free your mind of impure thoughts and look at this picture. If you did not understand that article, you probably understand what RAID is when you see this:


No Comments

Quantum notebook soon to be available…

There’s a great vision of the future when there are quantum computers available:

“Processor: 6.8GHZ CPU (AtomChip┬« Quantum┬« II processor / System Compliance: Two Operating Systems with Voice Command (Microsoft┬« Windows┬« XP Professional and Linux┬«) / Memory: 1TB Quantum-Optical non-volatile RAM (NvIOpSRAM-SODIMM 200-pin) / Storage: 2TB non-volatile Quantum RAM (NvIOpRAM-ATA IDE) / Optical Drive: DVD Super Multi / LCD Display: 12.1ÔÇŁ WXGA (1280 x 800, 16:10) TFT Glare Type LCD display with 1.3 Mega pixel CMOS camera / Video & Graphics: Two Integrated graphic controllers [Intel┬«855GME internal graphics, support Intel┬« DVMT (Dynamic Video Memory Technology) and AtomChip┬« DVM (Dynamic Video Memory)] / Communication: 10/100 Base-T LAN on board, MDC Fax/Modem V.90/V.92 on Board, 802.11a/802.11b/802.11g WLAN, Intel┬« PRO/Wireless 2100/2200BG/2915ABG network connection, WiFi, Bluetoth, GPRS -with Bluetoth antena, CMOS camera with USB interface, Mega pixel resolution CMOS image / Pointing Device: Synaptics touchpad with 4 way scrolling button / Application Launch Key: E-mail, Internet, Capture, WLAN / LED Status Indicator: Power, Suspend/Resume, Battery Charging Status, Quantum Storage Access, Num Lock mode, Caps Lock mode, Scroll Lock mode, WLAN Lock mode / Keyboard: 3.0mm travel, inverted-T, 88keys with 2 windows key (Internet & Microsoft For Connectivity) / Interface Ports Front Side: One 4-in-1 card reader slot (support SDIO/SD/MS Pro/MS), Audio line out, Stereo Microphone-in / Interface Ports Left Side: LAN port, Modem port, SVGAÔÇôout port, One Type II PCMCIA slot (support CardBus), 1394A port (mini jack) / Interface Ports Right Side: USB 2.0 ports x 3 / Interface Port Rear Side: DC-in, Kensington Lock / Audio: Built-in two stereo speakers and Built-in Microphone / AC Adapter (Input: 100-240V AC, 50-60Hz, 1.5A. Output: 20V DC, 2.5A, 50W) / Battery: 6 cell Li-lon battery pack. Battery Charge: 3 hrs charge time to 100% capacity while system off and 4 hrs charge time while system on. Battery Life: Approximately 8 hours for AtomChip┬« Quantum┬« II processor and 3 hours for 4 x Intel┬« Pentium┬« M processors 1.7CHz / BIOS: AMI, Support PnP, password, Bootable from USB and DVD / Power Management: ACPI 2.0 compliance / Smart Battery System Support / Security: Kensington Lock / Size: 320.0(W) x 242.0(D) x 22.0(H)mm/28.0mm (front/back) / Weight: 1.9kg(when fully equipped with AtomChip┬« Quantum┬« II processor) and 2,20kg(when fully equipped with 4 x Intel┬« Pentium┬« M processors) / Packing Accessories: Quick Guide, Support CD (Driver, Utility, Manual), AC Adapter, Power Cord, Battery Pack, BOSE Headphone Music System with noise Cancelling.”

The processor.

the non volatile RAM…


No Comments

benchmarking the storages: iSCSI Enterprise Target (+Software RAID0 3 drives)

This time we got a Software RAID0 Volume with 3 drives which made a volume size of 1465176195 sectors = 715418 Megabyte.

Interface-Transferrate with a blocksize of 128 sectors at 0.0 percent of the capacity:

sequential read rate medium (unthrottled): 78407 Kilobyte/s
sequential read rate Read-Ahead (Latency 0.90 ms): 61569 Kilobyte/s
repeatedly sequential read (“coretest”): 67066 Kilobyte/s

permanent transfer rate: (blocksize: 128 sectors):

  • Average: 75843.9 Kilobyte/s
  • Minimum: 69892.3 Kilobyte/s
  • Maximum: 78817.2 Kilobyte/s


  • Average: 30138.9 Kilobyte/s
  • Minimum: 29723.3 Kilobyte/s
  • Maximum: 30658.8 Kilobyte/s

access time read:

  • Average: 12.95 ms
  • Minimum: 0.16 ms
  • Maximum: 26.16 ms

access time write:

  • Average: 14.78 ms
  • Minimum: 0.34 ms
  • Maximum: 44.73 ms

access time read (<504 MByte):

  • Average: 5.87 ms
  • Minimum: 0.11 ms
  • Maximum: 19.22 ms

access time write (<504 MByte):

  • Average: 6.80 ms
  • Minimum: 0.29 ms
  • Maximum: 25.17 ms

No Comments

benchmarking the storages: Promise VTrak m500i RAID0

This time we got a Hardware RAID0 Volume with 4 drives which made a volume size of 2431630530 sectors = 1187320 Megabyte.

Interface-Transferrate with a blocksize of 128 sectors at 0.0 percent of the capacity:

sequential read rate medium (unthrottled): 46290 Kilobyte/s
sequential read rate Read-Ahead (Latency 1.52 ms): 44762 Kilobyte/s
repeatedly sequential read (“coretest”): 67861 Kilobyte/s

permanent transfer rate: (blocksize: 128 sectors):

  • Average: 46935.6 Kilobyte/s
  • Minimum: 38778.7 Kilobyte/s
  • Maximum: 47382.7 Kilobyte/s


  • Average: 33474.6 Kilobyte/s
  • Minimum: 32815.1 Kilobyte/s
  • Maximum: 33758.3 Kilobyte/s

access time read:

  • Average: 12.80 ms
  • Minimum: 3.65 ms
  • Maximum: 38.18 ms

access time write:

  • Average: 0.96 ms
  • Minimum: 0.47 ms
  • Maximum: 16.54 ms

access time read (<504 MByte):

  • Average: 6.60 ms
  • Minimum: 0.68 ms
  • Maximum: 1103.38 ms

access time write (<504 MByte):

  • Average: 0.61 ms
  • Minimum: 0.47 ms
  • Maximum: 3.24 ms

No Comments

strange picture resizing when you have a high-DPI display

At the moment I am working on a Dell Latitude D800 notebook which is equiped with one of those nifty 1080i widescreen displays. That means: 1920×1080 in a 15 inch TFT display.

I got the machine preconfigured – and noticed that the Internet Explorer resizes every picture – Firefox doesn’t:

You can clearly see the difference

So I searched and found the solution for the problem: It looks like someone set the DPI Settings in the Display Settings to 125% – which caused Internet Explorer to resize every picture by 25%… what a stupid behaviour…If I would like to let Windows resize anything it would be the font size… not the size of the pictures that cannot be resized without quality loss:

No Comments

benchmarking the storages: iSCSI Enterprise Target (+Software RAID0)

This time we got a Software RAID0 Volume with 2 drives which made a volume size of 976784130 sectors = 476945 Megabyte. This Volume is exported with the iSCSI Enterprise Target Software for Linux.

You may ask: Why only 2 drives this time? – The answer: The guy who did the setup of the linux machine did not know how to delete the RAID5 volume he created for the previous benchmark. So there where only 2 250 Gigabyte drives left…and yes: we sometimes have a subliminal feeling that we have to hurt him.

Interface-Transferrate with a blocksize of 128 sectors at 0.0 percent of the capacity:

sequential read rate medium (unthrottled): 65399 Kilobyte/s
sequential read rate Read-Ahead (Latency 1.08 ms): 67237 Kilobyte/s
repeatedly sequential read (“coretest”): 55854 Kilobyte/s

permanent transfer rate: (blocksize: 128 sectors):

  • Average: 54098.5 Kilobyte/s
  • Minimum: 51692.8 Kilobyte/s
  • Maximum: 54530.8 Kilobyte/s


  • Average: 30086.6 Kilobyte/s
  • Minimum: 29700.0 Kilobyte/s
  • Maximum: 30895.6 Kilobyte/s

access time read:

  • Average: 13.61 ms
  • Minimum: 0.11 ms
  • Maximum: 26.78 ms

access time write:

  • Average: 15.70 ms
  • Minimum: 0.36 ms
  • Maximum: 41.20 ms

access time read (<504 MByte):

  • Average: 6.22 ms
  • Minimum: 0.11 ms
  • Maximum: 21.31 ms

access time write (<504 MByte):

  • Average: 7.31 ms
  • Minimum: 0.29 ms
  • Maximum: 26.66 ms

No Comments

benchmarking the storages: iSCSI Enterprise Target (+Software RAID5)

This time we got a Software RAID5 Volume with 3 drives which made a volume size of 976784130 sectors = 476945 Megabyte. This Volume is exported with the iSCSI Enterprise Target Software for Linux.

Interface-Transferrate with a blocksize of 128 sectors at 0.0 percent of the capacity:

sequential read rate medium (unthrottled): 61384 Kilobyte/s
sequential read rate Read-Ahead (Latency 1.15 ms): 67472 Kilobyte/s
repeatedly sequential read (“coretest”): 54294 Kilobyte/s

permanent transfer rate: (blocksize: 128 sectors):

  • Average: 51913.4 Kilobyte/s
  • Minimum: 49738.2 Kilobyte/s
  • Maximum: 63889.7 Kilobyte/s


  • Average: 9080.2 Kilobyte/s
  • Minimum: 6650.9 Kilobyte/s
  • Maximum: 10129.3 Kilobyte/s

access time read:

  • Average: 13.47 ms
  • Minimum: 0.12 ms
  • Maximum: 28.65 ms

access time write:

  • Average: 38.82 ms
  • Minimum: 10.10 ms
  • Maximum: 108.19 ms

access time read (<504 MByte):

  • Average: 6.09 ms
  • Minimum: 0.12 ms
  • Maximum: 19.58 ms

access time write (<504 MByte):

  • Average: 14.42 ms
  • Minimum: 0.37 ms
  • Maximum: 75.51 ms

No Comments

benchmarking the storages: Promise VTrak m500i

After all strange things that we got to deal with using the Promise VTrak we made some benchmarks (which took about 10 hours each complete run).

So here are the results of the first test:

We made a Hardware RAID5 Volume with 5 drives which made a volume size of 1945310850 sectors = 949859 Megabyte.

Here are the results in all their beauty:

Interface-Transferrate with a blocksize of 128 sectors at 0.0 percent of the capacity:

sequential read rate medium (unthrottled): 43570 Kilobyte/s
sequential read rate Read-Ahead (Latency 1.62 ms): 39447 Kilobyte/s
repeatedly sequential read (“coretest”): 68155 Kilobyte/s

permanent transfer rate: (blocksize: 128 sectors):

  • Average: 44896.2 Kilobyte/s
  • Minimum: 18416.7 Kilobyte/s
  • Maximum: 45645.9 Kilobyte/s


  • Average: 29821.7 Kilobyte/s
  • Minimum: 12688.0 Kilobyte/s
  • Maximum: 30792.4 Kilobyte/s

access time read:

  • Average: 13.58 ms
  • Minimum: 4.13 ms
  • Maximum: 104.33 ms

access time write:

  • Average: 9.58 ms
  • Minimum: 0.48 ms
  • Maximum: 2930.85 ms

access time read (<504 MByte):

  • Average: 6.14 ms
  • Minimum: 0.48 ms
  • Maximum: 37.94 ms

access time write (<504 MByte):

  • Average: 3.74 ms
  • Minimum: 0.47 ms
  • Maximum: 1258.26 ms

No Comments

a lot storage space (not a Promise Storage Appliance)

So today it continues. After the dissappointing things that happened around the Promise VTrak m500i we started to put the other machines together to do some testing. Therefore we have:

Machine 1:

  • Pentium 4 (Prescott) 3 Ghz

  • 2 Gigabyte of RAM

  • onboard Promise(!) FastTrack 597 SATA

  • 4 Port Onboard SATA

  • 2×80 Gigabyte (Western Digital)

  • 3×250 Gigabyte (Western Digital)

  • 1xGigabit Ethernet (Broadcom)

  • Windows Server 2003 Enterprise Edition

Test-Setup (Machine 2 is not in this picture…)

the drives and behind the fans…LOUD!

a hot-swap fan and SATA cable close-up shot…

Machine 2:

  • Pentium 4 (Prescott) 3 Ghz

  • 2 Gigabyte of RAM

  • 2xPromise(!) FastTrack S150 TX4 SATA

  • 7×250 Gigabyte (Western Digital)

  • 1xGigabit Ethernet (Intel)

  • Linux 2.6.something

Machine 2 on top of Machine 1…

Machine 1 is destined to be the Documentation-Server running a Sharepoint Portal Server 2003. But we’ll do some testing with String Bean WinTarget 2.0 iSCSI Target Software. On Machine 2 the iSCSI Enterprise Target will provide access via iSCSI.

Our testing would be this: We setup a RAID5 on each machine and export this 500 Gigabyte Volume as an iSCSI Target Volume. We then benchmark this volume. After that we do this again with a RAID0 across the 3 drives. (so 750 Gigabyte Volume).

At the moment we only test with Windows on the client-side. We’re using h2benchw for this purpose.

Stay tuned for the results. Oh I almost forgot: The Promise VTrak m500i successfully synched a 1 Terabyte iSCSI Volume and I was able to mount and format this volume. The first test showed that about 66 Megabyte/s are possible on this RAID5 volume. The benchmark is running since about 2 hours – would take 6 more hours to complete. Of course we’ll also will test a stripe on the VTrak m500i.

Source 1: FastTrack S150 TX4 SATA
Source 2: String Bean WinTarget
Source 3: iSCSI Enterprise Target
Source 4: h2benchw

No Comments

Promise VTrak M500i Review and News…

I wrote a short article about the lately arrived Promise m500i Storage Array and it’s not-functioning.

After we had to wait the complete weekend it was time to call the technical support again and ask for purification…well sort of :-).

The good news: The guy remembered that he talked to me. He even remembered what we talked about. That’s really not common sense when it comes to technical support. So this is a thumb-up for the support.

He sort of apologized for mistakenly kidding me on friday…but don’t be vengeful.

So he told me that there is a BETA firmware for our brand new m500i that is supposed to fix all the problems we have (at least the ones he remembered). The only thing that he needs from me beforehand was a signing under a disclaimer he would send me. Sure! Show me the disclaimer and I’ll tell you if I would sign it.

And he showed me:

Supplier is providing this Beta version Product to Customer without charge and at Customer’s specific request. Customer understands and acknowledges that this Beta version Product has not been fully tested by Supplier. This Product is provided to Customer “as is” and with Customer assuming all risk of use of the Product.

Except where prohibited by law, Supplier DISCLAIMS any and all warranties, express or implied, by statute or otherwise, regarding the Products including,without limitation, any warranties for fitness for any purpose, quality, merchantability, non-infringement, or otherwise, and any warranties arising out of a course of dealing, trade usage, or trade practice. Supplier makes no warranty or representation concerning the suitability of any Product for use with any other item. Customer assumes full responsibility for selecting Products and for ensuring that the Products selected are compatible and appropriate for use with other goods with which they will be used. Customer assumes and accepts all risk associated with procuring and using a Beta version product.

Supplier DOES NOT WARRANT that this Product is free from errors or that it will interface without any problems with purchaser’s components or computer system. It is the responsibility of the purchaser or end-user to back up its computer or otherwise save important data before installing any Product and to continue to back-up its important data regularly.

Supplier shall not be liable for the cost of procuring substitute goods or services, lost profits, unrealized savings, equipment damage, or for any other general, special, consequential, indirect, incidental, orpunitive damages, whether in contract, tort, or otherwise, notwithstanding the failure of the essential purpose of the foregoing remedy and notwithstanding that Supplier has been advised of the possibility of such damages.

Date___________ Signed ___________________

The text formatting was not made by me – it’s the original formatting of the disclaimer. Of course I told the technical-support guy that I am not willing to sign this disclaimer. To loose the software warranty in any way is not what I would consider an alternative for a brand-new 6000 Euro device.

He understood my consideration about that issue – even though he tried to convince me to sign it. I didn’t. He then wanted to call me back after he checked the alternatives with his … whoever he was talking to… he called back 10 minutes later.

The alternative would be to deal with the bugs. He even seemed to know now and finally why the array does not sync successfully: I tried to make a 3 Terabyte Logical Drive. With my firmware version the m500i only supports a maximum size of 2 Terabyte per Logical Drive. Okay that’s at least some kind of solution. So I configured a 1 Terabyte Logical Drive for testing and voil


create your own iSCSI Target (with Linux…)

You don’t know what an iSCSI Target is? Well simply an iSCSI Target is the “device” that represents a logical disk volume on an iSCSI Server. If you don’t know what iSCSI is – please read here and come back.

So let’s take the case that you would like to make a harddisk volume available across the network. The best block oriented protocol that you could use (and the best sounding…BUZZWOOOOOOORD!) is iSCSI. So there are some commercial iSCSI Implementations for Windows and there are some iSCSI Implementations for Linux. Today cosrahn implemented a solution based on the Linux iSCSI Enterprise Target Project.

This article will give you a list of things you have to do to get it to work by yourself:

You would need:

  1. a linux server with a kernel version 2.6.11 or later

  2. a free disk volume/partition/block device (even a file)

  3. a direct network connection to the machine that would like to use the iSCSI Target

    1. alternatively you have a iSCSI enabled network switch…quite expensive…

  4. a linux or windows machine that has an iSCSI Initiator installed

The iSCSI Initiator is the client part of the solution. It can either be software (like the Microsoft iSCSI Initiator) or hardware (like the Adaptec 7211C). In our case it’s a complete software solution.

Cosrahn installed a Gentoo Linux with a kernel. That took the most of the setup-time. Then he downloaded the source from the iSCSI Enterprise Target page and compiled and installed it.

Then he edited the config file in /etc/ietd.conf:

# iscsi target configuration

# the password NEEDs to be minimum 12 characters to be standard conform
IncomingUser username longpassword

# Targets definitions start with “Target” and the target name.
# The target name must be a globally unique name, the iSCSI
# standard defines the “iSCSI Qualified Name” as follows:
# iqn.yyyy-mm.[:identifier]
# “yyyy-mm” is the date at which the domain is valid and the identifier
# is freely selectable. For further details please check the iSCSI spec.


# Users, who can access this target
# (no users means anyone can access the target)

IncomingUser username specialtargetpassword

# Logical Unit definition
# You must define one logical unit at least.
# Block devices, regular files, LVM, and RAID can be offered
# to the initiators as a block device.
# in our case /dev/sda4 is a partition without any filesystem.

Lun 0 Path=/dev/sda4,Type=fileio

# Alias name for this target
Alias iSCSI-Test-Volume

# various iSCSI parameters
MaxConnections 1
InitialR2T Yes
ImmediateData Yes
MaxRecvDataSegmentLength 8192
MaxBurstLength 262144
FirstBurstLength 65536
DefaultTime2Wait 2
DefaultTime2Retain 20
MaxOutstandingR2T 8
DataPDUInOrder Yes
DataSequenceInOrder Yes
ErrorRecoveryLevel 0
#HeaderDigest CRC32C,None
#DataDigest CRC32C,None

After that we started the iSCSI Target daemon and configured the Windows Client to use the target.

This is the “startup-screen” of the Microsoft iSCSI Initiator.
You can see the node name of the iSCSI client.

This is the “iSCSI Server Discovery” Tab of the Initiator.
This is where you enter the “username” and the “longpassword”

Now there’s the Target Tab of the iSCSI Initiator. If you entered the Logon-Information to the iSCSI Server correctly you should see the target that are exported by the iSCSI Enterprise Target daemon. In our case there are two targets available.(despite the configuration above just exports one…just ignore it, you can create as much as you like)

You have the choice to let the iSCSI Initiator reconnect to the Target automatically on boot time and of course you have to authenticate yourself to the target. In this case it’s “username” and “specialtargetpassword”.

After you entered everything correctly the “inactive” flag would become a “connected” flag. Now fire up the Disk Management in the Computer Management MMC-SnapIn.

Voil├í! You should now see a Disk with the size of the exported target…Just start using it like any other hard disk. Create a partition and a filesystem.

Simply as it could be. That’s it. There are some more options like the Header/Data digest – which is an option to ensure data integrity. While this surely will increase data security it also will increase the CPU usage when you access the iSCSI drive. In our test a 1.7 Ghz Pentium-M CPU had an utilization of about 35-40 percent whilst the iSCSI server only had about 7 percent. We think that the client has so much utilization because of the quite cheap network interface card (realtek…). In fact the utilization is quite massive if you enable Header/Data digest-

Source 1:
Source 2: iSCSI Enterprise Target
Source 3: Microsoft iSCSI Initiator
Source 4: Adaptec 7211C (iSCSI HBA Copper)

No Comments

the Promise VTrak M500i arrived: another 3 terabytes (soon?) to be available

Do you remember when I wrote about the brand-new Promise M500i iSCSI appliance? Yeah. They delivered it four days ago. And we’re nearly there getting it to work…but as usual: first things first!

This is the little bugger:

And that’s the whole rack:

So what happened:

We ordered the Promise M500i and 15 250 Gbyte drives:

Promise M-Class | Physical Drives Summary
PdId Model              CfgCapacity Location       OpStatus   ConfigStatus

1    HDT722525DLA380    231.90GB    Encl 1 Slot 1  OK         Dedicated Spare
2    HDT722525DLA380    231.90GB    Encl 1 Slot 2  OK         Unconfigured
3    HDT722525DLA380    231.90GB    Encl 1 Slot 3  OK         Unconfigured
4    HDT722525DLA380    231.90GB    Encl 1 Slot 4  OK         Unconfigured
5    HDT722525DLA380    231.90GB    Encl 1 Slot 5  OK         Unconfigured
6    HDT722525DLA380    231.90GB    Encl 1 Slot 6  OK         Unconfigured
7    HDT722525DLA380    231.90GB    Encl 1 Slot 7  OK         Unconfigured
8    HDT722525DLA380    231.90GB    Encl 1 Slot 8  OK         Unconfigured
9    HDT722525DLA380    231.90GB    Encl 1 Slot 9  OK         Unconfigured
10   HDT722525DLA380    231.90GB    Encl 1 Slot 10 OK         Unconfigured
11   HDT722525DLA380    231.90GB    Encl 1 Slot 11 OK         Unconfigured
12   HDT722525DLA380    231.90GB    Encl 1 Slot 12 OK         Unconfigured
13   HDT722525DLA380    231.90GB    Encl 1 Slot 13 OK         Unconfigured
14   HDT722525DLA380    231.90GB    Encl 1 Slot 14 OK         Unconfigured
15   HDT722525DLA380    231.90GB    Encl 1 Slot 15 OK         Unconfigured

 Global Physical Drives Settings
 Return to Previous Menu

(that’s the serial console of the M500i)

Source 1:

And after getting the drives into the bays and the appliance we started configuring the device. As usual: one hot-spare and the rest for RAID50 (later we configured it for RAID5 only). The machine started the initialization/synching process…after several hours the 60 percent mark was reached…we let it ran the night and the next morning…”Synching: 20%” was on the screen…WTF?! we thought. Apparently no power failure occurred that night. The machine just restarted the sync-process…and as if this wasn’t bad enough it kept restarting the process on and on…

We reconfigured the device for RAID5 and started again…but after another day the same things happened: restart after restart…In fact there never was an error of any kind – neither in the logs nor in the activities display.

We even figured out another (maybe) bug of our M500i: When we unplugged it completely from the power the device forgot some of it’s settings (like the network configuration…). And the more serious bug: When I try to set the Management Network Interface to DHCP the whole administration consoles lock up completely. That means: no serial console or web interface until we rebooted the device (that behavior is reproducible).

And as this would not be enough the device booted with this text on the serial console:

IBL_ROM Ver 1.05.0000.00
Built 18:18:10, Jun  9 2005
Warning: Detected BAD NVRAM Signature
Engineering mode = Active
SDRAM Memory Size ...................... 256 MBytes
Initialize chipset for ECC(8) Memory ... Done
Initialize OS reserved memory .......... Done
Initialize Disk reserved memory ........ Failed
Platform Setup completed.
+fis load kernel
go 0x1008000
Load & Run:  IBL_RAM
DHCP Disabled
IP:, Gateway:
Default server:, DNS server IP:
IBL_RAM Ver 1.05.0000.00
Built 18:19:24, Jun  9 2005
VTrak M500i
Copyright 2005, Promise Technology, Inc.
== Executing boot script in 5.000 seconds - enter ^C to abort
IBL_RAM> fis load kernel
IBL_RAM> go 0x1008000
Loading software modules .......

Interesting isn’t it?

I called the technical support…but as usual it happened what seems to be common sense for technical support: No help at all. They told me that the M500i restarts the sync process probably because there are no iSCSI clients connected yet…I was told to add a client at least to give the M500i a chance to sync. (Actually this advice did not and does not make any sense at all…but I tried)

After this advice did not solve any problem (as expected by me) I again called the technical support. Nobody was there to talk to me in a language I understand so they offered me in english to call me back when there’s someone available. They called back half an hour later, told me something about a new firmware and that they (of course) never have heard of any problems like that with the M500i – and that they would have to ask the guys in the states if they are allowed to send me a new firmware (for a brand new device that does not work at all…). Another theory was that there might be a hardware error (this is my preferred theory)…but they would like to try the software update first…(the update would fix the DHCP problem hopefully). So they told me to call back at 1700 – and when I called back the answering machine told me that I call outside of business hours…

I really hope that this was a mistake and not intended…expect more “Stories of interest” when they will answer my desperate phone-call tries…

Source 1:
Source 2: solving data storage problems: How long can 3.5 Terabyte last?

1 Comment