Good point about the RSSI vs Signal to Noise Ratio. It's a little bit of a tangent (OK. A lot of a tangent), but if you wanted to measure the actual packet throughput, you could also do a UDP blaster-blastee. The way this works, is you have a UDP sender that blasts out UDP packets. Inside the packet you put a sequence number and a very high resolution time stamp. On the receiver, check that the sequence numbers come over in order, and count the gaps. Also, subtract the current time stamp from the previous one to measure the inter-arrival time. Drop the inter-arrival times into an array of buckets (like a histogram: 5 packets were between 2.0 and 2.5 msec, 8 packets were between 2.51 and 3.0 msec, etc), and you can do stats on the amount of jitter you got, and also count out of order packets, dropped packets, etc.
Generally, you would want to have the slower machine blast to the faster machine. Try both wired, one wired, one not, both wireless to see what you get. Sorry I can't send the source code. I wrote it a few years ago when I worked for WindRiver (VxWorks) and it's copyright them. Also, it was written for VxWorks not Windows, so the code would be a little different anyway. Or, if you don't want a programming exercise, you can just use a sniffer and look for the same information. There are some free ones out there. And, on UNIX variants you should be able to do an etherfind. Luke P.S. Sorry about the HTML earlier. Eudora sneaked that one in on me. At 03:41 AM 9/24/2002, David Neves wrote: >Another word of caution to add to this discussion, a >better signal strength (signal power) _doesn't_ >necessarily translate to better signal quality... >You might be getting a signal with more power but also >with more interference, or, in a worse case scenario, >you might be getting mostly interference on the same >band than the signal. > >I'm not really sure what the meter is indicating, but >if it's only RSSI (signal strength) it's not so >surprising what you're getting... There is usually >another indication that some drivers supply (at least >on Windows XP and WinCE) that better relates to the >actual "Link Quality" (usually goes with this name). >You probably should refer to this indication to better >estimate the performance of the wireless link. > >As Luke said, your problem might also be related to the >platform you're using, rather than the actual wireless >link (radio environment). A possible source for your >trouble might be the USB device driver... Not that the >USB should be a significant bottleneck as this bus' >performance achieves 12Mbps (unless you have large >bandwidth devices on the bus like USB hard-disks or >CD-Recorders) but a buggy device driver could cause >a great degradation of performance... As a first >approach you should check for updated drivers! >(really!) :-) > >=---------------------------------===----------------------------------= >David >Neves -<|>- ><mailto:[EMAIL PROTECTED]>[EMAIL PROTECTED] >----- Original Message ----- >From: <mailto:[EMAIL PROTECTED]>Luke >To: <mailto:[EMAIL PROTECTED]>Raymond Sinatra Jr. >Cc: <mailto:[EMAIL PROTECTED]>[EMAIL PROTECTED] >Sent: Monday, September 23, 2002 8:03 PM >Subject: Re: [BAWUG] Linksys Wireless USB Adapter Troubles > >Surfing speed is directly affected by CPU and real memory available. You >don't say if you have equivalent CPU's, but I'm guessing you don't. My >surfing speed went up a lot when I upgraded. > >USB probably doesn't help. It's another layer of bus and drivers to go >through, and has a much lower B/W than pcmcia or PCI. The drivers, in XP's >case, might be quite significant. > >Also, your PC may not be correctly setup in the TCP/IP parameters. Try the >performance tests at dslreports.com. Specifically look for the TCP/IP >(MTU) configuration wizard. > >Signal strength has quite a lot to do with antenna orientation and >position relative to concrete walls, metal boxes (computer case, fridge, >oven). Also, remember that radio waves bounce around, and can form peaks >and valleys in intensity. Closer to the router will not necessarily give >you better results. Also, his meter might just be calibrated differently >than yours. If he's getting good performance, I wouldn't worry about what >the meter says. But, you can move the box around the room and between >rooms to see if the power meter changes. > >Lastly, attach your box to the network with a wire and compare the >performance results to see if it is your computer or the wireless connection. > >Luke > >At 09:39 AM 9/23/2002, Raymond Sinatra Jr. wrote: >>Okay, frustrated wireless guy here. My roommate has a laptop with Windows >>XP and a wireless network card. I have a desktop with Windows XP and a >>USB adapter. My roommate and I are roughly the same distance from the >>router (his room is 10 feet further away from the router). All equipment >>is Linksys. His signal strength is never better than 'good' and >>oftentimes stands at 'low'. My signal strenght is 95% of the time listed >>as 'excellent'. He is currently surfing at much higher speeds, most times >>more than twice the speed. What gives? Does signal strength have anything >>to do with the speed of surfing? Is the fact that the adapter is >>connected via USB slowing me down? >> >>Please help... >> >> >> >> >> >>Do you Yahoo!? >>New <http://rd.yahoo.com/evt=1207/*http://sbc.yahoo.com/>DSL Internet >>Access from SBC & Yahoo! -- general wireless list, a bawug thing <http://www.bawug.org/> [un]subscribe: http://lists.bawug.org/mailman/listinfo/wireless
