Researchers unlock fiber optic connection 1.2 million times faster than broadband

LINK: Researchers unlock fiber optic connection 1.2 million times faster than broadband | Popular Science (popsci.com)

This is quite amazing. Read that again folks. 301 terabits per second. And no new fiber optics need be installed to handle it. Just need the equipment designed to receive/send at such speeds. Gamers would be in heaven that's for sure. Activists and politicians? Well...just depends on how you look at it I guess. Will probably be a love/hate relationship.

 

IT stuff is always beyond me, so I have no idea what difference that will make for my mundane purposes.

 

I am an old timer, I remember paying extra for 600 bit/sec in Compuserve.

 

Don't know what you consider "mundane" so can't really tell you if it will help you or not. But for gamers who download lots of games this will turn a 20 minute download or a 5 hour download into a half a second download. For those that stream online, high quality vids only limited by the equipment you're using (IE: no more lag). For those that upload video's to sites like Youtube it will take a split second, only wait time will be for Youtube to render it (make it compatible for the software they use) themselves. For businesses/corporations it will allow for faster transmission of large amounts of data...and as they say in the business biz...time is money. Same applies on the government level with the transmission of data...only for them it could be time saved in reacting to various situations. And for anyone that has had to wait for downloading an update for Window's et al that will take a split second also (probably the only thing that would affect you if none of the above applies).

Of course I can see a downside to this...specifically making it easier for hackers to...well...hack as a faster internet speed does help them out.

 

Yeah, I'd imagine this will cost quite the penny at the start, but once it gets widespread the price will level out to something manageable for the common person.

 

I’m sure it will drive the use of AI, and expand the services and capabilities of just about everything online.

 

Man, this will make streaming JAV a breeze! LOL

 

This is exciting, for sure.. However, the only real negative I can see is that this will move the AI needle vastly farther than our current standards, as processing speeds are somewhat throttled ATM. Hive AI working at 301 terabits per second is the only thing that sends shivers down my spine.

 

So, it occurs to me that "instant everything" has always seemed to many in the tech field as something virtuous. But think about the limitations then of humans to consume "instant everything".. it isn't possible. So, given that there isn't a single use case for it that supports human interactive experience, what is the over arching goal of it? To make machines sentient? Why?

 

I can see reasons for concern.

We can’t seem to manage social media now. And it is the sandbox compared to AI.

 

That's a great step forward. I imagine all the routers, switches, and modems would need to be upgraded too.

 

I'll try to put some context around this, please let me know if I need to expand on something in greater detail so it can make sense.

This type of technology is generally used by ISPs when connecting to other ISPs or to extend their backbones, not consumers. Additionally, higher bandwidth throughput does not necessarily mean lower latency. This type of technology has existed for a very long time. Effectively what this means is that they've opened up another wavelength that can be used within a fiber cable. We can already do things with DWDM which basically means utilizing multiple wavelengths over a single cable to aggregate throughput.

When talking about obtaining 300 Tbps, we also have to think about the infrastructure that can support that, as well as the cost. The networking device, whether it be a switch or a router, would be extremely costly. Additionally the optical transceiver required to terminate the connection would also be extremely costly. The processing needed to manage that would be extremely expensive, and then, if you plan on doing any network policy enforcement, you're adding a ton of overhead. An example of the cost for just the transceiver would be demonstrating the price of a current 800G transceiver. A single transceiver is ~$5-8K, and you would need at least two to make the connection.

As far as "gamers" would be concerned, they primarily care about latency. Games themselves, outside of the downloading example, use very little bandwidth when actually playing the game. Copper ethernet and wireless will continue to be the medium of choice within households because fiber is too delicate, and transceivers/fiber switches are too expensive. The typical ethernet speeds are 1G and 10G. However, NBASE-T interfaces can use speeds within these (2.5G and 5G).

Long story short, breakthroughs in ISP interconnection speeds helps boost speeds per households as ISPs refresh their equipment every 5-10 years. You would need these speeds on every hop between you and your content in order to see the benefit; otherwise, you'll still be bottlenecked by the slowest device in your chain. Additionally, you're still going to be limited by the speeds of the interfaces on your wired computers or the wireless access points/network interface cards. You'll also have to factor in the capability of the hosts to read and write data at those speeds.

The value of these speeds won't be realized until quite some time in the future. There's been a lot of research and development into fiber transceivers over the past decade and we continue to see incremental growth as the need/value for these speeds is realized.

 

It won't be available. The current fiber-optics came out long time ago (in 1996), and they haven't even made that available for a large number of customers. It requires physical work to run the cables to each household.

 

As a former cable TV tech ops manager, I always believed it was bandwidth that gave us speeds, not the frequency being used to transmit data. For example, the forward part of the internet (that's that part coming at you) is contained in a series of digital channels 6 MHz wide. The return (the part leaving your computer) is also contained in a series of digital channels 6MHz wide, but they are fewer in number than the forward channels. All the forward channels utilize something called channel bonding, so let's say there are 6 forward channels, that would mean the bandwidth is 36MHz wide which can carry a ton of data real fast. Same thing with the return.

 

Researchers unlock fiber optic connection 1.2 million times faster than broadband

Heck I would be thrilled to have broadband out here in the country. It is all a matter of perspective.

 

You should read the article linked in the OP.

 

Oh, ok. So they can use the existing fibers. Then it must be the signal which changed.

 

We just got an Apple Vision Pro and these types of speeds would be required for extreme high definition immersive video.

This may also help speed up automation processes that are becoming more and more commonplace.

 

I'm no expert in this area or specific tech but that blindingly fast. Traders/markets are likely the first to be in line for this tech and I can see how AI will be improved/advanced by this. Interesting times ahead.....

 

Might even be compression with integrity.

 

For example 8 terabits equals 1 terabyte.
1,000 gigabytes equal 1 terabyte.

At 242 megabytes per second (United States internet speed) it would take over 10 hours to download 1 terabyte
At 301 terabits per second it would take a fraction of a second to download 1 terabyte


This is absolutely huge!

 

A few things to consider. As I mentioned above, I spent most of my adult life working in cable TV technical operations and I don't know of any modem capable of going past 1 gig - even then, 40 to 60 megs is held back for what they call house keeping. (so if you sign up for 1 gig service the best you can get is 960 megs) Because of the time it would take to develop modems to operate at such high speeds, along with the time it would take to put them in people's homes, these high speeds won't be a consumer reality for a while.

Second, think about this: If a person can download a terabyte in a faction of a second, what does that data get stored until it's sent to the hard drive? Almost certainly the device doing the downloading would need at least a terabyte of RAM to store the downloaded stuff before it's written on the hard drive. Every device inside a computer would need to move data as fast as these download speeds.

These high speeds will require a lot of hardware upgrades at the user end. At first these speeds might be used by government, universities and large corporations, but trickle down to consumers over time - lots of time.

 

Huh? This doesn't make any sense.

 

Which part?

High speeds will be needed for immersive video streaming as it is significantly larger than regular video streaming.

 

Modern automated traing works on low latency. The Exchange is in Equinox, NJ and financial traders in NYC have direct feeds of all real-time trades sent to their offices via a direct connection. Several years ago the feed exceeded 10 Gbps, but they're nowhere near the throughput requirements of current networks. Where the low latency portion comes in is with the automated trading. They analyze the feed that comes in from Equinox and they automate trades based on custom algorithms. The idea being if Stock A, B, and C go down, but Stock D goes up, they know that the next stocks impacted will be E and F and they can make microtransactions in real time. Additionally, they monitor flight/shipping logs and other analysis as part of their secret sauce. It's a low latency game trying to beat their competition to the trade to maximize their gains or minimize their losses.

When it comes to individual traders making decisions they don't need anything remotely close to the automated trading business.