With Cox cable, my tests show my download speed to be around 7200 kbs (that’s bits). Still confused, but does this indicate 7.2 Mbs? If not, do I have to divide the 7200 by 8, then convert?
Or what?
Is this speed good, bad or indifferent?
That’s what it is and it is a very good speed. The people that named connection speeds should be shot because they are way too easily confused with file sizes expressed in kilobytes and megabytes.
Yes, you have 7.2 Mbps, which is megabits per second. However, all files in common nomenclature are expressed in bytes, so if you are downloading a file that’s 7.2 MB (megabytes,) it will take eight seconds, not one.
Still, though, 7.2 Mbps is pretty damn good. Cable internet has a caveat, though. It’s maximum speed is almost always faster than DSL, however, DSL is dedicated per customer (I have a 3.0 Mbps DSL, and I get 3.0 Mbps most of the time.) The cable’s bandwidth is "shared’ by a number of households, so that 7.2 might drop down to a 2.0 at peak times if there are lots of households signed up in your “block.”
Thanks, guys, for the information. Guess I’m lucky, as it does seem pretty fast, but then anything would seem fast after dialup!
I live in the boonies, so probably not more than one or two other houses around here signed on to Cox. About the lowest speed I’ve encountered in occasional tests is 6700 or so, highest 7300.
No DSL up here, so picked cable, and now i learn that it is faster anyhoo. Now and then things work for the best.
I’ve found that I have to run several tests in a row to get a meaningful number. I just ran the CNet one a couple of times with about 40% difference from the highest to lowest. Average was about 1100kb/s here in Comcast cabletown.
Sometimes the shared nature of cable connections can work in your favor. I am subscribed to a 3000kb/s service via Comcast yet consistently have a download speed of 10000kb/s (my upload speed does not change though). With my old DSL connection, I subscribed to 1500kb/s and never achieved 1kb/s more than that.
A point to remember is that link speed is measured in terms of total throughput, ie all bytes in all packets. TCP/IP has significant packet overhead, such that sending a 1MB file may well push 1.2MB of actual bytes onto the wire.
So really a 10/1 conversion factor is a better real-world approximation. 7.2 megabits per second of bandwidth will move about 7.2 megabytes of files per second.
Communications engineers have been using “bits per second” for many decades. It’s the correct unit for data transfer rates, along with the symbol. The bit and second are fundamental base units. The byte is a non-standard unit invented by computer engineers. It is usually, but not always, 8 bits. The proper term for an 8-bit byte is octet, which is commonly used in standards documents. A kilobit is always 10^3 bits, not 2^10 bits. A megabit is always 10^6 bits, not 2^20 bits. A gigabit is alway 10^9 bits, not 2^30 bits. This is standard usage for SI prefixes.
Data transfer rates should always be specified as “bits per second” or “symbols per second” (baud).
I don’t know the answer to your question, but I have one of my own. Is your email extension cox.com or cox.net or other? The reason I ask is a friend of mine has cox as his provider and I’ve been trying to get ahold of him.
Thanks.
For me it is like a thermostat. Who cares what it says, it is the result that counts. I put a stop watch to download times and compare apples to apples on the same site, server, time of day, same browser, etc.
My old computer, 600 Hz CPU and a satellite connection with latency, will sometimes beat an 1.7 Ghz machine on local comcast cable because I go with the right browser for the server, the right time of day, etc.
I keep a big BMP picture up on AOHell space that I know the URL to by heart and so when I visit folks, stores, etc., I punch that in and time it. I know how fast I can get it on average and that gives me a reasonable idea of the speed I am looking at on a day to day real world basis.
The gages do not always truthfully indicate what is really happening in the real world.
That only checks the speed of the route to your picture.
When the data request leaves your computer it goes up the pipe (or non-truck like tube) to your providers routers. From there, depending on the provider, it might go to another set of routers in the same building before the request heads towards your bmp file. Then, when it arrives at the AOL building it might jump around routers in there as well. The hops and route your request takes can vary greatly between providers.
One provider might take a 27 hop path to the bmp and seem slugish. Another might only take 17 hops and seem quite speedy. A visit to google.com might take both 34 hops.
Case in point. I work for a company that provides an online service. One element has to do with video. Our product stack, and most our customers, are on the west coast. We have a few servers on the east coast.
For the video element it is best if the client has a quick, short connection to the video server. We found some clients were getting choppy video. Testing the routes we found our provider on the west coast was adding a bunch of extra hops just routing the traffic to our main routers. We switched our video server to the east coast and it cut our clients hops down by 25-35%.
Most our clients could take a taxi to our main product stack but they got better video if we route that traffic across the country.
The internet is like beer, it’s all about the hops.
Although I never use the Cox email (it is terribly clunky), it is @cox.net.
Are you sure about that? DSL (usually ADSL but there’s also SDSL) is subject to a contention ratio, and I’ve seen that vary from 20:1 to 50:1 which means that you’re sharing that 3 Mbps (for example) bandwidth with 20 to 50 other people. Indeed the last time I checked, some ISPs reserve the right to put you on a high-contention line if you use too much bandwidth. So if you’re on a 3 Mbps line with 50:1 contention and everyone is troffing, then everyone will get 60Kbps. But chances are you’re not all troffing at the same time, so everyone seems to get full bandwidth.
And the water.
What? Who mentioned beer, anyway?
I might be wrong, but doesn’t ADSL contention normally indicate a maximum contention, which is unlikely to be actual ratio for most ADSL users?
I seem to have phrased that badly: the ADSL line is shared between N people, but not all of those N people will be using it all at once. Sometimes there will be more people using it so your effective bandwidth goes down; sometimes there will be fewer people using it, so you’ll seem to get full speed.
And don’t forget that the ISP itself has a maximum outbound / inbound bandwidth, shared across all its customers.