The Fastest Internet Ever

To better conceptualize how fast this new fast internet is, let’s put it in perspective. The speeds discussed by the researchers are along the lines of one terabit per second. One terabit is equivalent to 122,070,312.5 Kb. What does that mean? Well, according to an online conversion tool, what this new chip downloads in one second would take my DSL connection at home 11 days to download or a 56Kb dial-up connection 206 days to download. Sounds like a big difference to me.

Would this chip need to be purchased and plug into each individual computer or would this chip be on the side of the server and internet providers? Is this just research results or is there a company already in the works to implement this? I want to know more about this! This kind of speed boggles my mind

c-rod87: No. Since the chip seems to be using fibre-optic cable (which is incredibly expensive to make, however it is very common everywhere amongst internet and phone line providers because of it's incredible data transmission capacity). You could likely get the chip put in but I doubt the internet provider is going to provide your house with fibre optic cable.

Also, we'll probably see this technology put into LANs first and then into ISPs later. Seems it is at the moment research results without a company working to implement this; but as the Australian government is supposedly working to improve the communications infrastructure over here we'll probably see it soon - howerver; going by past experience the technology we develop here will be implemented here last, lol.

One little problem with this is that the infrastructures will likely still be using phone line cables to homes and business and the like so the speed is limited by the capabilities of this low level technology - but as a nice little tool to reduce network load for LANs and WANs this little chip looks to be the shiz-net.

Hold on a bit, you need more than the flux capacitor chip.

This only applies to a chip that can help handle data rates in fiber optic networks. It could be a huge benefit to allowing the servers to be able to handle a lot more clients, but it won't help the data coming to you via DSL, satellite, etc.

Also, the data can only be sent as rapidly as it can be sourced -- i.e. somewhere there are RAMS, disks and the like providing the data for transmission.

It is a good deal, and if it really works, it should be impressive. But, there are a few more fish to fry before you can download 10 DVDs in a second

You can't discount the fact that this is would be the first market release of a chip utilizing such technology. It's about time we step away from the copper.

You can't discount the fact that this is would be the first market release of a chip utilizing such technology. It's about time we step away from the copper.

The issue is not transmission speed, but processing speed for traffic. In current applications of transmitted data (including copper wire transmission) the limiting factor is the processing of the data. In order for a signal to reach its destination, routers have to examine packets of information and send it. This is a relatively slow process. The *real* reason that this is exciting is because the Internet as a whole will get faster. I have heard discussion that if consumers continue to get faster connections to the Internet it could potentially overload the routers that must process the information. Not only does faster processing mean that this is unlikely, but it also means that less equipment will be needed to do the job. This is likely to make Internet access cheaper. This new technology means that the speed of the Internet can continue to scale as it has been and it won't break the infrastructure.

This might be research now, but you can believe if it is viable technology it won't be sitting around for long. What I would like to know is what effect if any will this have on processor speeds? What are the prospects of such an integrated circuit for use in a general purpose processor? If it can speed up the processing of signals in a network, can it be adapted for use as a general purpose processor? I for one wouldn't mind a graphics card that is 60 times faster than what I have now.

Another issue however is the limit in internet addresses. The Internet continues to grow in speed but also in size. At some point we will have to take the plunge and switch to IP version 6 or some other solution.

I don't understand your initial comments:
"Who wants to live forever?" and "Or grant robots the same rights as humans?"

Are you just being flippant?
Who wants to die? Are you in the habit of avoiding medical attention so that you will die sooner? When will you start?

As to robots, better be careful there. If we aren't "nice" to them, they may do more than just ignoring us, once their intelligence surpasses ours. They may return our hostility.

Can someone explain this:
The original article says the switch can change states in about a picosecond.
A picosecond is 1/1,000,000,000,000 (1E-12) of a second, isn't it?
then why is it, the switch can only make 1 million switches in a second? How many state changes does it require to switch the path of the signal? A million?

That seems wrong...

This seldom gets mentioned but the dirty little secret is that bandwidth for most ordinary web sites is controlled by how much the site wants to pay for it. Every webhost has a pay-per-gigabyte per second plan. Notice how even though you often can download large files from selected sources at speeds around 500 to 1000 kbytes/second for a typical cable connection (those sites who pay the likes of Akamai to handle such feeds) the actual loading of most sites is often limited to a couple of hundred kbytes per second at most.

It gets worse for sites that are dynamically constructed, namely just about any media outlet, e.g. the New York Times, or anyone using a content management system, since the server has to build the page and then the web server has to send it out. That all takes a lot of horsepower and that's where the real bottlenecks are. On top of which sites with heavy advertising have to fetch that third party stuff and send it along too.

Vast improvements in network delivery speed will really be felt for the those downloading huge files, or using streaming video from Akamai, et al.