Then do a simple test. Set two computers to transfer a very large file and maybe some small app running to simulate normal work. Set the 100 cards in and transfer the file and measure. Do the same with the 1000 cards. How much different? Try that same thing with a cpu intensive task. Now how much are they different? People have tested this before. "Best Practices", Event viewer, host file, perfmon, antivirus, anti-spyware, Live CD's, backups, are in my top 10

0

right i'm gonna do that test right now. NOT I don't have two computer with 100 and 1000 and a 100 and 1000 switch/router. In fact I don't have a router or switch period, wow so that is why I'm asking for advice about them. I'm gonna stick with what Curt R said. You seem to say many strange things. such as http://www.computing.net/answers/ne...

0

Ok, suits me fine. Lemme know how it turns out. "Best Practices", Event viewer, host file, perfmon, antivirus, anti-spyware, Live CD's, backups, are in my top 10

0

Then do a simple test. Set two computers to transfer a very large file and maybe some small app running to simulate normal work. Set the 100 cards in and transfer the file and measure. Do the same with the 1000 cards. How much different? We've got some issues going on here at work that required me to do some data transfer testing yesterday. I quickly built a 1 GB file to transfer on my work PC and then transfered it to our department laptop via a Cisco 2950 24 port switch in my office which is a 100 Mbps switch. The transfer took 2 minutes and 30 seconds. I then went down and did a transfer from laptop to PC in another department. The laptop was plugged into the same Baystack 5510 (1000 Mbps) switch. The transfer took 40 seconds. I then repeated the transfer again with the file on a server (on our server subnet) transfering it from the server to a client (on the client subnet). I repeated that transfer once for each client subnet in our main building here (that's 7 client subnets). Each transfer took between 45 seconds (fastest) to 1 min 15 sec (slowest). It's worth noting that in this case, unlike the above transfer that took 40 seconds, the data had to traverse a firewall/router. The above 40 second transfer was from one port to another on the same switch (no firewalls to traverse). Just to be extra sure that the 'backplane' of the computer didn't affect transfer rates (note, above transfers were all done on P4 computers) I did a transfer from the UNIX box in my office to the laptop via a 1000 Mbps switch. That transfer took 43 seconds. The UNIX box is a Pentium III running at 800 MHz with 128 MB's of RAM. A 3 second difference between the PIII and the PIV's. Try that same thing with a cpu intensive task. Now how much are they different? With the CPU running at 100% on something else, yes, it will take it longer to process the incoming data. No matter how highend your NIC is, the CPU still has to recombine the packets into their original form. Anybody with a half ounce of brains understands that with the CPU loaded doing something else, yes, it will take longer for the CPU to process the incoming data. But how long it takes the CPU to process the data has nothing to do with how fast it transfers across the network.

0

Thanks for that post Curt R It is informative and nice to know 1000 is faster. (I can't help laughing at that statement) I crunched some numbers since I was bored. assuming your file was 1024 mb 100 mbit lan 1024 mb in 150 s 6.28 mb/s theoretical max lan speed 12.5 mb/s percent of max speed used 50.24% 1000 mbit lan 1024 mb in 40 s 25.6 mb/s theoretical max lan speed 125 mb/s percent of max speed used 20.48% (stupid thing doesn't keep spaces in formatting, had a nice side by side chart like thing.) I wouldn't take those numbers for much other than that the 1000 mbit was faster than the 100 mbit lan. When I get a 1000 mbit switch or router and nic for my extra computer I'll test this myself.

0

Your numbers came out to pretty much what we calculated. Under ideal conditions one could probably get better transfer rates but for the most part one has to remember the listed transfer rates are theoretical and no matter what you do, you're not going to get the full 100/1000. The main point being what you already figured out, it took less time to transfer the same amount of data across the 1000 Mbps than it did on the 100....period! I'm redoing the tests today with a larger file (5.6 GB's) and am extending it to closets throughout the building (more hops away from the core switches). So far we're getting the same throughput we did yesterday. So what does all this mean? It means, if your clients have 1000 Mbps NIC's, and you have need to move large chunks of data frequently, it pays you to get a 1000 Mbps switch.

0

That is cool my numbers came out the same. A couple days ago I read about a new Netgear wireless n router that is 70$ and was quite excited, only to find out it has 100 (edit was 1000) mbit lan. I was shocked, the wireless is faster than the wired. I guess they cut costs to make a cheap wireless n router. Back to the topic of this thread. I'm thinking about getting a LINKSYS WRT54GL which is 50$ for the wireless router and the cheapest 5-8 port 1000 mbit switch I can find that has good reviews. The TRENDnet TEG-S5 has good reviews and is only 28$. I guess that will fit my needs and budget, hopefully prices will drop by Christmas season. if anyone has input on the stuff i've listed please let me know

0

A couple days ago I read about a new Netgear wireless n router that is 70$ and was quite excited, only to find out it has 1000 mbit lan. I was shocked, the wireless is faster than the wired. I guess they cut costs to make a cheap wireless n router. You must have misread or misunderstood. Wireless N is 100 Mbps capable. You stated above the router is 1000 Mbps. So, you have the hard wired connections being 10 times faster than the wireless. Keep in mind, 802.11n is not yet a standard. What this means to the end user is, it's highly unlikely you'll have any interoperability between brand names. As long as you stick with one OEM's equipment, it should work ok. But mixing and matching brands would likely be problematic. Also, once 11n is standardized, you may suddenly find your equipment does not work at all with any other brand names due to it being manufactured before the 11n was standardized and now it doesn't meet the standards. Just something to consider. Myself, I would hold off on buying any 11n devices until it has been standardized. Back to the topic of this thread. I'm thinking about getting a LINKSYS WRT54GL which is 50$ for the wireless router and the cheapest 5-8 port 1000 mbit switch I can find that has good reviews. I don't see anything wrong with either device. LinkSys is pretty reliable and now owned by Cisco. I've had some issues with a LinkSys wireless Access Point, but that's not likely to be an issue in a small home environment and you're looking at a router, not an access point. The best thing you can do is read reviews. That's what I always do before buying. Customer reviews can tell you a lot.

0

well i made a mistake saying 1000 mbit in my previous post, which i corrected and noted. That wireless router has 300 mbit wireless and 100 mbit lan speeds. Also wireless n is up to 300 mbit, so I'm confused by your first paragraph other than saying I messed up somewhere. Which i did. I don't plan on using the wireless at all myself, so as long as it does b/g I'm fine. I just have some friends with laptops so it'd be nice, and how common are wired routers? My personal philosophy on wireless is it's great for internet and streaming some stuff, but for real data transfers I'd plug a laptop into the wired network to get higher speeds.

0