Highest MB/s transfer rate on a home LAN?

Page 4 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Status
Not open for further replies.
Trance: I would sugest you take a good long look at this drive. thats about the same price I payed for my 250 GIG around a year or more ago. Denser platters usually equates into faster access speeds, look at the advertised seek times too, they are close to that of the Raptors.
 
lol. i knew i should have just pulled out the calculator. sorry. i meant 1min, lol. u know, regular Gbit might not be that bad afterall. But if it was possible for faster performance around the same budget then why not right.
 
Lol, i just read all the reviews on newegg about that drive, because the last one sort of concerned me, the second review to the last states that the drive performs better than his 'noisy 74 GB Raptor' LOL

Pros: Quiet and alot more performance than i had from my noisy 74gb raptor.

Cons: Took 2 hours to format, seems kinda slow to format for its size.

Other Thoughts: 5 year warrenty is nice, i guess time will only tell if it will outlast that warrenty.

I get the feeling this person is exaggerating somewhat . . .
 
lol. u think buying the 2 cheaper 250s for 80 bucks and Raiding them will be better?

Well, these are relitively new drives, with denser platters, this usually means the drives are faster, because the effect is simular to increasing the RPM of the spindle. IF you want to be absolutly sure, and you can wait, wait for some reviews from toms, or anandtech to pop up, and see what they say( i think madwand said something about a HDD reviewer page a few pages back).

Making a guess based on past experience, reading the reviewers comments from newegg ( the one guy claiming difficult installation is an idiot . . .), I'd have to say the extra 60 bux for these two drives would have to be worth it, thats IF you can afford the extra 60 bux . . . if not, then you got to do what you got to do. As for the Samsung drives you mentioned earlier, I have no dirrect experience with thier drives, so i couldnt recommend either way.

Ah something i nearly forgot, Im guessing your system is SATAI capable, not SATAII capable, and I'm unsure if this would effect performance of these drives at all. I would have to say very little, and the denser platters would STILL play into your favor at the least.
 
cool, looks like i'll be doing some hunting for reviews on these babys. other than that, yes i do believe they can be worth the extra sixty. i'll probably get one, then bench it, and if satisfied get another. thanks dude.
 
cool, looks like i'll be doing some hunting for reviews on these babys. other than that, yes i do believe they can be worth the extra sixty. i'll probably get one, then bench it, and if satisfied get another. thanks dude.

The main problem i have with recommending older stuff is 'futureproofing' Obviously, the larger, and more modern technology of a HDD, is going to be better for future proofing.

Another consideration, try using froogle, to find a better price on the drives, maybe you can save a few more dollars, but dont try saving too much, some companies have a really crappy RMA policy, and as you can see by the last guys review, Seagate seems to have slipped in thier return policy, and is doing refurbished replacements. If this is true, they are going to lose favor, from me included, and if they continue along this path, they may find all thier loyal customers, migrating to other manufactuers. My motherboard manufactuer of choice (ABIT) is starting to pull this, and if they continue, I ay find myself doing what Ive said Id never do again, and thats buy from Asus . . .

Newegg, and other like e sellers, have a pretty good RMA policy though, so no biggy ( for now) :)
 
Edit: I read the whole thing and realized why the title exists, lol. but its still a good article.

hahaha, well i had that problem once with my dogs but i put them both in the garage. They are puppy boxers, both pretty white coat, boy and girl. Had them for a week so far. but hey when u get a chance check this out, i found it not too long ago. Looks scrumptious. lol.

http://www.networkworld.com/news/2004/101104infports.html

So do u think i can get a 4 port or 2 port NIC and achieve 2 or 4 Gbit somehow. Or does this stick with the strictly Linux policy?

http://www.newegg.com/Product/Product.asp?Item=N82E16833106217
http://www.newegg.com/Product/Product.asp?Item=N82E16833106209

This might help too: http://inteleval.ententeweb.com/Lit/products/252454-
005US.pdf
 
On pg.3 of the NetworkWorld article, you can see exactly the issue that I've brought up here:

"Aggregating multiple Gigabits is great," says Nathaniel Mendoza, an IT technician at the SDSC. "But you can't get a stream over one Gigabit through," on the trunked connections. Although the server had a total of about five to seven Gigabits of bandwidth, no single flow, such as a file transfer, could exceed 1G bit/sec.

So the problem is not the hardware, but software / protocol which allows > 1 Gb/s single connection/stream. So far, it looks like Linux has the only solution.
 
yeah, the hell with technology. What they should do is after they make a product, they should test it using real world benchmarks. Then name it by its real world performance. For example: They should call Gbit Ethernet 600Mbit Ethernet, not 1000Mbit. But oh well, they like the attractive names.

I'll just have to settle for Gbit, and wait for 10Gbit to become affordable. Hopefully the article is right about 10Gbit getting as cheap as 400 per module within the next year or so.
 
Hey trance, here is a review of the 750GB, and 500 GB variants of the same drive I linked you to last night, and the day before. If you read the whole article, you can see that the 150 GB Raptor spanked the crap out the the Seagates in alot of tests, and the Seagates Spanked the crap out of the Raptors in a few tests. Then, once you read the real world end results, you'll see the 'stale taste' end result I was speaking of. While game level loading difference may be drastic in a Game or two, the real world times are all basicly really close. But the interresting results, is that two of those 750GB Seagates in RAID 0 OMPLETELY spanked the crap out of a single 150GB Raptor in the Anti virus scanning test. right now, for a single vs single drive to have for shear performance, it seems the Raptor is the drive to have (I would have guessed this already), and its even faster than its older brother, the 74GB version (I also would have guessed this)

Anyhow read the review, read for yourself :)
 
yeah, the hell with technology. What they should do is after they make a product, they should test it using real world benchmarks. Then name it by its real world performance. For example: They should call Gbit Ethernet 600Mbit Ethernet, not 1000Mbit. But oh well, they like the attractive names.

I'll just have to settle for Gbit, and wait for 10Gbit to become affordable. Hopefully the article is right about 10Gbit getting as cheap as 400 per module within the next year or so.

Typicly, the results alot of people get, 1GbE should be called 480Mbit 😉 Since USB2 has a theoretical throughput speed of 60MB/s . . . But anyhow, the end speed you get from networks is tied to alot of different factors, motherboards ability to actually provide enough data to the port, the interface use (PCI would obviously hold the back seat to PCIE IF, this was the bottleneck), the media being used, can play a factor, System I/O, even the OS can play a major factor, alot of people claim that compared to Unix/ Linux, windows has an inferrior TCP/IP stack, and in some cases, I've even proven this myself. Dont forget patch cables, switch etc can also play a factor, as well as cable length . . .
 
yeah, the hell with technology. What they should do is after they make a product, they should test it using real world benchmarks. Then name it by its real world performance. For example: They should call Gbit Ethernet 600Mbit Ethernet, not 1000Mbit. But oh well, they like the attractive names.

I'll just have to settle for Gbit, and wait for 10Gbit to become affordable. Hopefully the article is right about 10Gbit getting as cheap as 400 per module within the next year or so.

C'mon, I think you're just getting greedy. Note that gigE became affordable not long ago, and it's basically dirt cheap right now. Also note that say 90% of corporate environments don't do gigE to the desktop. So your home is going to be way faster than most of these. Also note that even corporations find it hard to justify 10 GbE except in very limited areas with tons of traffic (i.e. users, not single users by a long shot).

We also have a long long way to go before HD's come close to meeting anywhere near 10 GbE performance. We'll probably also need faster processors and other components to match.

Consumer gigE CAN do nearly full speed. I've benched 900 Mb/s TCP throughput using consumer gear. Add TCP overhead, and this is pretty close to max gigE capability. This is just amazing, and some corporate folks still don't believe it, and are shaking their heads at the deals we're getting considering how hard they had to work & pay for gigE not long ago.

The teaming solutions may or may not come in the near future. Until then, a 3x performance improvement for peanuts is a great deal.
 
I have a 1Gb lan and a D-Link 4100 gaming router. It's fast but I don't know how to monitor the optimum transfer rate.

There are some notes on what you can do on pg. 1.

Heheh, me being a hobbyist programmer, I just wrote a simple VB.NET WMI perform tool 😉
 
yeah ur right. well i'm just gonna have to do alot of research, which i have done most of it, and choose the best performing products for the money.
 
well if u think about it, imagine not being limited to ur cables, switches, length, motherboard, NICs, or anything else. With 10Gbit all it would hopefully come down to is how many hard drives you can gather up to pump as much output as possible.
 
well if u think about it, imagine not being limited to ur cables, switches, length, motherboard, NICs, or anything else. With 10Gbit all it would hopefully come down to is how many hard drives you can gather up to pump as much output as possible.

Things don't generally work this way. When you introduce a very high-speed component into a system, everything else becomes a bottleneck. 10 GbE has strong cabling requirements, CPU requirements, bus requirements, switch requirements, etc., etc. The HD is already a severe bottleneck at 1 GbE. In current cheap NAS boxes, the CPU/ASIC is a severe bottleneck, and you can't even get 1/2 GbE performance, regardless of the number and speed of drives and efficiency of the outside network.

If you're not getting at least say 30 MB/s file transfer performance over consumer gigabit, the odds are that 10 GbE wouldn't help you either -- something other than the network is a bottleneck. Of course, it's possible to have an under-performing 1 GbE network, but in my limited experience, it's very likely that something other than the network is the bottleneck in such cases. The steps above 30 MB/s are pretty much exposing all these limitations, starting with drive performance. At the higher end of gigabit, when you're approaching 70-80 MB/s, then yes, the 1 Gb/s network limit comes into play. But until then, you should probably look elsewhere.

PCATTCP can tell you what your TCP bandwidth is. Drive to drive copy in a single computer can tell you what your drive bandwidth is. These are good starting tools for such analysis.
 
Things don't generally work this way. When you introduce a very high-speed component into a system, everything else becomes a bottleneck. 10 GbE has strong cabling requirements, CPU requirements, bus requirements, switch requirements, etc., etc. The HD is already a severe bottleneck at 1 GbE. In current cheap NAS boxes, the CPU/ASIC is a severe bottleneck, and you can't even get 1/2 GbE performance, regardless of the number and speed of drives and efficiency of the outside network.

If you're not getting at least say 30 MB/s file transfer performance over consumer gigabit, the odds are that 10 GbE wouldn't help you either -- something other than the network is a bottleneck. Of course, it's possible to have an under-performing 1 GbE network, but in my limited experience, it's very likely that something other than the network is the bottleneck in such cases. The steps above 30 MB/s are pretty much exposing all these limitations, starting with drive performance. At the higher end of gigabit, when you're approaching 70-80 MB/s, then yes, the 1 Gb/s network limit comes into play. But until then, you should probably look elsewhere.

PCATTCP can tell you what your TCP bandwidth is. Drive to drive copy in a single computer can tell you what your drive bandwidth is. These are good starting tools for such analysis.

I agree with this somewhat, but 'we' are forgetting about somethigns. Lets take USB 2.0 for example, the transfer limitation is 60MB/s but you often see devices operating at 5 MB/s, so whats the problem? Its the controller thats the bottleneck in this case most of the time, and the same problem can exist in a 1GbE network situation.

This is why I suggest, keeping cables as short as possible, using properly terminated cables, using a decent performing switch (doesnt have to cost an arm and a leg), network cards can sometimes play a big factor in this as well, but with the good performing Intel NICs costing $20 now ( i think madwand said) there is no reason for this to be much of an issue. HDD wise, as long as you have a well performing drive (or in this case we're talking two in RAID0) this SHOULDNT be too much of an issue, since two of those seagate 320's should give you ATLEAST 90MB/s throuput Remember, we're going to be reading from the stripe here, not writting for outgoing packets), and should be more than enough to feed the GbE line. Lastly, we have protocol, and TCP/IP stack (remind me if im forgetting something), Jumbo frames, I'm not too familiar with, so I cant really speak from experience reguarding this.

I cant say what other people expect from a GbE connection, but me personally, I'd say if you're getting less than 60MB/s, then you're doing something wrong.
 
cool, so lets wait and see. do u think they will come out with something like 5gbit for midrange performance.

Id have to say no. Judging from past 802.x specifications jumps (atleast wired networks) they seem to be advancing <current tech> x10 (powers of ten ?) *brain is numb * been working out in the sun all day (112 F !)
 
cool, so lets wait and see. do u think they will come out with something like 5gbit for midrange performance.

Id have to say no. Judging from past 802.x specifications jumps (atleast wired networks) they seem to be advancing <current tech> x10 (powers of ten ?) *brain is numb * been working out in the sun all day (112 F !)

lol, do your thing. is there any simple program that can monitor your transfer rate and all your PCs on a network?
 
Status
Not open for further replies.

TRENDING THREADS