Allowing for the fact that the Internet isn't always "up to speed," I'm still mystified by why it doesn't turn on like electricity, as I'd expect.

Almost invariably, when I initiate a d/l it kicks in at a relatively high speed, quickly works its way up towards my max, and then plummets and works its way back up again, but in agonizingly slow increments of 1 or 2 Kbps, sometimes even 1 or 2 Cbps.

There's apparently some sort of resistance factor in effect, and I'd appreciate somebody's explaining what and where it is.

Thanks.

I'll give it a try....

Think of your access to the Internet through the computer as if it were a small soda straw connection. Through this soda straw flows all the information which, when assembled by your computer, is then presented on your screen as a web page.

Now I know what you're thinking....this is easy, it's just like water flowing through a hose.....more pressure and it flows with more authority or volume, bigger hose and it flows easier, open the faucet wider and you get more sooner, speed of light, etc --- and if we were simply talking about DC electricity, these analogies might be helpful in understanding what's going on with this process.

Now the bad news......it's not that simple -- instead of water we have to visualize the flow of sand through our soda straw. Each grain of sand is the same size, but slightly different in shape; containing both information (data) and instructions (overhead for handling) so that it can be analyzed then processed on arrival and reassembled, bit-by-bit (pun intended) once received......EVERY step of the way during the transmission of our data, there is intermediate processing taking place. Each grain of sand (data packet) is meticulously checked on arrival at each step for integrity and, if there is the slightest issue, it is asked to be re-sent through the soda straw.....once in the direct link between stations, our grain of sand will travel at the speed of light, but it will be sharing the soda straw with millions of it's closest friends while traveling from each intermediate origin to each next intermediate destination -- occasionally there will be minor collisions which interfere with the smooth flow of traffic, and so there will be diverts and re-routes along the way: sometimes it's just easier to turn left in Albuquerque and pass through Calgary Canada on the journey from Los Angeles to Denver......sometimes it just gets lost, and we have to recognize that we are missing a specific grain of sand and go back to the source to request a do-over re-send.....

...and this process above takes place silently in the background millions of times each time we send a request to the server from our computer asking to see a particular web page. The server may be next door or halfway around the world.....the path for our request may be direct with very few intermediate jumps, or it could go the long way around depending on network traffic issues....the response from the server might be immediate or it might be queued up for action in order with other actions already in-process.....once the individual grains of sand are all on your computer after their journey, they still need to be sorted, re-organized, moved through the various busses and caches and memory stations (stacks and registers and buffers, oh-my!) within your deuced Mac(hina) to get presented in human-readable format on your screen.

In the simplest terms, all the above says is that when you think about it, we have a pretty amazing thing happening when we surf the internet.....the fact that there are variations in response speed should not be a surprise. The fact that we can now move so much more data so much faster is more a factor of improved processing along the way than anything else.....carrying information by modulating light in optic cables rather than modulating audible tones over a pair of copper phone wires.

Beautiful!
Only reinforces my wonder that any of this ever works correctly.

Bravo! What a great way to explain it.

Wonderful explanation! (And I wholeheartedly concur with Ira's wonderment that the whole darn thing works at all.)

Not to detract from your analogy, though, I'll make my own (which I'm able to do only by virtue of yours), which I think is slightly clearer as respects my question.

My initial high-speed burst of data is the few people right in front of the doors when they open on "Black Friday."

Once they're in, the flow of customers slows down to a crawl while the people behind them in the crowd try to jostle their way through the doors.

And eventually, the store guards force the crowd into an orderly line, and the store quickly fills up with, well...shoppers is a polite way of putting it.

Many thanks!

It could also be the result of your ISP throttling your throughput.

If that were the case, why would I always get up to full speed, albeit slowly?

(I'm contracted for 7Mb, and I'm getting pretty close to that.)

Let's not forget that an important reason why your download speed varies over time is the multifactorial 'general internet condition'. Personally (~350KBps DSL), I've seen the (almost) immediate maxing out and then petering off phenomenon too, but more often I see the opposite, with download speed ramping up until finished or maxed out.

Occasionally, a particular download settles at a certain, sub-max speed. When that's really slow (or when I'm impatient), I might stop the download and resume it immediately. Quite frequently, things then take off and quickly hit max speed until finished. Doesn't always work, though, presumably reflecting source server issues. Whatever is going on, certain trends tend to hang around for days or weeks before settling on another pattern.