Unfortunately no one can be told what fun_plug is - you have to see it for yourself.
You are not logged in.
New DNS model number DNS-321 at http://www.dlink.com/products/?sec=0&pid=666
Looks just like the DSN-323 with a different front cover plate.
Also, the DNS-343 is listed at http://www.dlink.com/products/?pid=667&sec=0
Neither model is listed at the D-Link USA shop on-line store http://www.dlinkshop.com/default.asp .
Offline
Man that 343 looks nice!
Nice we can use Raid 5 now thats redundancy...
Last edited by nUll (2008-06-03 19:49:12)
Offline
The DNS-321 looks to have IP webcam support, could be a nice low end surveillance solution.
But is that the only difference for the 323?!
A better CPU and more Ram would be nice... and the case looks cool too.
We have to wait for the first reviews.
Last edited by assuncap (2008-06-03 20:02:02)
Offline
nUll wrote:
Man that 343 looks nice!
Nice we can use Raid 5 now thats redundancy...
RAID5 offers exactly the same level of redundancy as RAID1 - your data remains available if one (1) disk fails - if a second disk fails before the first has been replaced and the resync completed, you will still lose your data.
RAID5 has the both advantages and disadvantages when compared to RAID1 ....
The advantage is effiency (from a storage point of view) you lose one disk's worth of space, in a four disk system that would be 25% of the space lost, as compared to 50% for RAID1
The disadvantages are ...
a) greater complexity - in the case of a system failure it is going to be more difficult to retrieve the data, you will need to move the entire RAID array (all the disks) as compared to just one disk in a two disk RAID1 array.
b) write bottlenecks - all writes require a checksum to be calculated and written to the disk.
This last can become a serious issue in multi-user database applications (where I doubt you'll find the DNS-343) where individual records are being accessed randomly, to the extent that some developers with whom I have done implementations will point blank refuse to install on a RAID5 system.
c) read access takes a performance hit if operating with a failed disk - the data can no longer be read directly, but must be extracted from the checksum data.
It would be interesting to see which processor is being used in the DNS-343 and to see how the read/write performace compares to the DNS-323, I'm also wondering if the DNS-321 is going to replace the DNS-323 as we know it today.
Offline
Dns-343 :
all the dns-323 tricks(fun_plug 5, and all the f_p5 packages working), 128 MB ram inside, same Marvell proc. as in the 323,
oled is cool, two fan (4x4 cm) and 1 giga eth.
Offline
assuncap wrote:
The DNS-321 looks to have IP webcam support, could be a nice low end surveillance solution.
But is that the only difference for the 323?!
A better CPU and more Ram would be nice... and the case looks cool too.
We have to wait for the first reviews.
You can do webcam support with "compatible IP cameras" with the DNS-323, D-Link's DCS-950 & DCS-1110 can both record to any NAS, I don't know if you realize that all though the IP camera is shown in the sketch, it is not mentioned anywhere in the text - and the unit does have a 1GHz processor & 512 MB RAM.
Unfortunately the DNS-343 brochure does not mention processor or memory, but it does have UPS support, both shutdown & startup.
Offline
craft - what's the throughput like in a RAID5 configuration?
Offline
fordem wrote:
- and the unit does have a 1GHz processor & 512 MB RAM.
I think you are mistaken. The 1GHz processor & 512 MB RAM is what is required on your desktop computer, not the specs of the unit. I imagine it has identical specs as the 323, but it lacks the USB port. There is no mention of using it with a printer so I think the 321 name makes sense. It is a step down from the 323 from what I can see.
Offline
Looks like is been out in the UK for a while also in AU.
Offline
fordem wrote:
craft - what's the throughput like in a RAID5 configuration?
I'll test this tomorrow.
Right now I'm testing ext2 vs. ext3 system.
more details :
little change in the housing - horizontal HDD slots (instead of the 323/313 vertical slot(s) )
jumbo frame supported (like in DNS-323 1.05 fw)
Two buttons in the front - "power" and "next" - push power over 5 sec. ->shutdown the device
and push "next" -> next screen in the oled.
Offline
assuncap wrote:
The DNS-321 looks to have IP webcam support, could be a nice low end surveillance solution.
But is that the only difference for the 323?!
A better CPU and more Ram would be nice... and the case looks cool too.
We have to wait for the first reviews.
321 is less than 323 - so I think memory reduced (...32 mb) , minus USB port = low-cost version of the DNS-323
Offline
haldan wrote:
fordem wrote:
- and the unit does have a 1GHz processor & 512 MB RAM.
I think you are mistaken. The 1GHz processor & 512 MB RAM is what is required on your desktop computer, not the specs of the unit. I imagine it has identical specs as the 323, but it lacks the USB port. There is no mention of using it with a printer so I think the 321 name makes sense. It is a step down from the 323 from what I can see.
You're right - I was looking at a print of the data sheet and the inkjet was streaking at the end of the page - looking at the PDF it's very clear.
Offline
fordem wrote:
assuncap wrote:
The DNS-321 looks to have IP webcam support, could be a nice low end surveillance solution.
But is that the only difference for the 323?!
A better CPU and more Ram would be nice... and the case looks cool too.
We have to wait for the first reviews.You can do webcam support with "compatible IP cameras" with the DNS-323, D-Link's DCS-950 & DCS-1110 can both record to any NAS, I don't know if you realize that all though the IP camera is shown in the sketch, it is not mentioned anywhere in the text - and the unit does have a 1GHz processor & 512 MB RAM.
Unfortunately the DNS-343 brochure does not mention processor or memory, but it does have UPS support, both shutdown & startup.
well i had only a quick look on the diagram...
i don't know if the dns-323 does support ip cams out of the box or not. obviously it is a matter of software and i'm sure it supports it one way or another... that was not my point.
What matters for mainstream costumers, those that don't have any idea how things work, is what the box gives them on day 1. and that is what i was referring too. It may be marketing... but that small ip cam may sell them a few 1000 unites more.
Last edited by assuncap (2008-06-04 01:25:03)
Offline
craft wrote:
assuncap wrote:
The DNS-321 looks to have IP webcam support, could be a nice low end surveillance solution.
But is that the only difference for the 323?!
A better CPU and more Ram would be nice... and the case looks cool too.
We have to wait for the first reviews.321 is less than 323 - so I think memory reduced (...32 mb) , minus USB port = low-cost version of the DNS-323
i suspect that is true... it's very sad. Maybe what dlink wants is to compete with Linksys nas200's price tab.
men how i would love 256mb or 128mb on that machine.
I guess we need a dns-324 with oled screen(just to be cool) and 128mb of ram
Last edited by assuncap (2008-06-04 01:31:30)
Offline
assuncap wrote:
craft wrote:
assuncap wrote:
The DNS-321 looks to have IP webcam support, could be a nice low end surveillance solution.
But is that the only difference for the 323?!
A better CPU and more Ram would be nice... and the case looks cool too.
We have to wait for the first reviews.321 is less than 323 - so I think memory reduced (...32 mb) , minus USB port = low-cost version of the DNS-323
i suspect that is true... it's very sad. Maybe what dlink wants is to compete with Linksys nas200's price tab.
men how i would love 256mb or 128mb on that machine.
I guess we need a dns-324 with oled screen(just to be cool) and 128mb of ram
hmm in Singapore DNS-323's price is almost on par with linksys nas200 thou
Offline
nULL, you say the DNS-343 has been out in the UK for a while, but I can't find it for sale anywhere. Can you let me know where?
Cheers,
Gareth.
Last edited by gareth (2008-06-05 15:37:58)
Offline
craft wrote:
fordem wrote:
craft - what's the throughput like in a RAID5 configuration?
I'll test this tomorrow.
Right now I'm testing ext2 vs. ext3 system.
more details :
little change in the housing - horizontal HDD slots (instead of the 323/313 vertical slot(s) )
jumbo frame supported (like in DNS-323 1.05 fw)
Two buttons in the front - "power" and "next" - push power over 5 sec. ->shutdown the device
and push "next" -> next screen in the oled.
Craft,
Any performance numbers for us? Does this thing perform any better, or does it just have more disk capacity? Can you validate if the UPnP 4GB bug is in this firmware as well?
Curious peeps want to know.
Cheers!
bspvette
Offline
performance in raid5, that's what interests me the most too. cpu could be huge bottleneck here
raid6?
Offline
SilentException wrote:
performance in raid5, that's what interests me the most too. cpu could be huge bottleneck here
raid6?
Assuming the RAID6 question relates to performance degradation - if degradation occurs with RAID5, it will be worse with RAID6 - how much worse will depend on the technology used.
The DNS-323 is an OS driven RAID device, meaning that the RAID functions are handled by the operating system, which is fully aware that there are multiple physical disks, as compared to hardware RAID in which case the OS does not "see" the physical disks - if this technology is taken over to the DNS-343, there is likely to be significant performance degradation when using RAID5 due to the lack of a dedicated IOR processor, RAID6 would probably see even more degradation since twice as much checksum data would have to be calculated - even with hardware RAID, there is likely to be increased performance degradation over RAID5 due to the need to write the additional checksum data.
Offline
no actually, raid6? was the question if raid6 is possible. i do know how raid operates
Offline
fyi looks like the cmd line parse is busted
root@dlink-CDBC8B:/mnt/HD_a2/tmp/test# ./test -e 60
./test: line 1: syntax error: "(" unexpected
but what i see is interesting
memset(buf , 0 , sizeof(unsigned long)*2);
//erase the request block
memset(&request, 0, sizeof(request));
sock = socket(AF_INET, SOCK_DGRAM, 0);
if( sock > 0 )
strncpy(request.ifr_name, "egiga0", IFNAMSIZ - 1);
else
return;
request.ifr_data = (char *)buf;
buf[0] = 0;
buf[1] = cmd;
buf[2] = time;
status = ioctl(sock, (SIOCDEVPRIVATE), &request);
cant we make the same ioctl call using perl
or another scripting lan ?
Last edited by tc66 (2008-06-11 17:42:57)
Offline
The datasheet does not mention RAID6 - so depending on the RAID technology used maybe, maybe not. If it's OS driven RAID, I'm sure someone will find a way to manipulate mdadm and make it happen.
BUT - like RAID0 and JBOD - would there be any advantage?
In addition to the potential for performance degradation mentioned above, RAID6 would require two "drives worth" of storage space reducing the available capacity to 50%
For those not familiar with RAID here are a few nice links.
This one is an HP white paper comparing the perfomance of RAID + ADG (advanced data guarding), HP's name for RAID6, to RAID 0+1
This one is a chart comparing the, so to speak, cost of the different RAID methods - and if you're interested, there is a link to the entire "book" in pdf format in the upper left of the page.
For the record - I have no affiliation with HP/Compaq - I am HP APS certified and I do recognise that many of the innovations in the PC server market place are the result of their research, I was at one point very "pro Compaq" and have always been amused to see technology available on Compaq servers being introduced on Dell & IBM equipment many years later, (I am also Dell DCSE and IBM Business Server Support certified)
Last edited by fordem (2008-06-11 19:03:58)
Offline
raid 0+1 might be the best raid level to set up on dns-343.. i'm waiting for some test results and more important, the price tag in eu...
Offline
RAID 0+1 is also not mentioned on the website.
There have been some interesting debates on implementation of RAID 10, 0+1, 1+0 - the definitions of which vary from one manufacturer to another (a mirrored spanned array or a spanned mirrored array), some hardware based implementations dictate which can be used.
Depending on the implementation, the array may not survive the failure of a second disk - and this may also be dictated by the location of the second failure, relative to the first.
Offline
yup 'm aware. still, i'm quite confident that raid5 will be sucky on this cpu, as the copying files on dns-323 raid0 takes about 30% (it's jumpy). so alternative configurations should be considered
let's wait and see what options d-link gives us when someone gets and test this device.. it might be interesting to create two raid0 arrays or two raid1 (as two separate volumes) instead of 1+0 (or 0+1)
@tcc66: wrong thread but you can try it, the code should not be complicated when you have this as a background. if i find some time i'll look further into the source code i uploaded and try to fix the error..
Offline