Tech Thursday: The Faster Internet we Desperately Need

Gigaom reported yesterday that researchers at MIT have developed a method that could increase the speed of the internet 1,000-fold. (Wait - THAT'S A LOT! - you say? Why yes, yes it is!) Gigaom does a decent job of summarizing the original article, but if you'd like a bit more detail, and some extra commentary, you can find it at Physorg. Essentially, the team has devised a way to eliminate the need to convert optical signals into electrical signals in order to store data in short-term memory while waiting at a router. If that sounds complicated, you're probably right, and I''m not going to even try to describe how this will work. I will, however talk about why I think this is cool after the jump.

 First off,  the two (2) [I always find it amusing when articles do that!] key things to note are:
  1. Our internet will become much quicker with the new technology, while
  2. Incredibly reducing our energy consumption at the same time!
Traditionally, these two (2) [look! there it is again!] have had an inverse relationship, so it's astonishing to see that this group thinks they can now have a direct relationship. Although, one has to wonder that this is only the case because the current network so poorly designed that any change of this magnitude almost has to come with energy savings.

What does all this mean for you? Well, not too much in the short-run. Our overall usage of the internet isn't high enough, I believe, to truly see improvements in speed if we go this route. Additionally, the 1,000-fold increase in speed I believe is related to latency, and not actual bandwidth; and, since latency is already measured in milliseconds, I doubt we'd ever notice if latency ever dropped to microseconds. However, I do think this will help us in the long-run as we bring more and more users (especially heavy users) into the fold; lag times will surely increase as we move forward (imagine a household trying to stream multiple 3D-HD movies at the same time). Also, I think this type of network (opening up one-way 'highways' whenever a connection is made opposed to sending data via packets) is simply a better solution as more and more information is spread across longer and longer distances.


  1. So how long will it take a web page to load if it's one thousand (1000) times faster? Does it go from two (2) seconds to faster than my comprehension? From ten (10) seconds to faster than my comprehension? Can my computer react that fast?

  2. Eff you blogger. Thanks for deleting my comment.

  3. Well, latency does not deal with page loads - it's a measure of how long it takes to establish a connection; bandwidth (or, throughput), measures page loads. Both do factor into a user's perceived throughput (how fast is my internet?).

    My favorite example is AT&T's response to Verizon's "There's a map for that" campaign. AT&T's response is technically true: once connected, AT&T's throughput/download speeds is best in class. However, AT&T's notoriously shoddy connection performance leaves end users actually perceiving Verizon to have a faster network. It takes longer to actually secure a connection with AT&T; but, once connected, AT&T's connection speeds are awesome.

    Looping back to your original question, my best guess is that you won't experience much change if this method were to be implemented. However, if this change were made 10 years from now - after our internet consumption jumps incredibly - I'd expect end users to see significant improvements in overall 'internet speed'.


  4. So a high latency, high bandwidth connection will show a blank page for a few seconds then load quickly, while a low latency, low bandwidth connection will start loading right away but appear to do so more slowly? And a high latency, low bandwidth connection will cause me to shut down the browser, restart the computer, then go out and buy a new one?

    Is latency determined by the network (or the router or the modem)? Or does my computer determine the latency? Will I have a higher latency running multiple applications and browsers than if loading only a single page with nothing running in the background?

  5. Yeah, eff you blogger

    My comment wasn't deleted but I'm with you both in solidarity

    So if latency is 1,000x better does that mean the online games I play are going to have 0 lag, all the time? As far as I know latency is the factor there, not bandwidth, right?

  6. Correct on all your scenarios Aaron. Latency is network (which includes routers). Computers do not touch latency, as far as I know. Multiple apps using the internet will cause the bandwidth to be clogged and slow down your downloads.

    Royce - I do believe if latency is improved 1K-fold, online games will experience muuuuch less lag.