The seminal paper proposing the use of switched/routed interconnection networks on-chip (NoCs) was published by Dally and Towels 11 years ago in DAC'01: Route packets, not wires: On-chip interconnection networks. The idea of associating a router to each core and replicating it in "tiles" is not new either; Tilera was (IIRC) the first company to sell processors based on a tiled design, which was an evolution of the RAW research project. A related research project, the TRIPs, replicated functional units on each tile, rather than full cores. Intel has used a tiled design in the Polaris, SSC and MIC (which includes the forthcoming Knights Corner).

So no, the idea of using routed interconnects is not new at all. In fact, after reading the linked article, turns out that 2/3ths of the text are introducing the idea, and the last section details the contributions: Two ideas developed by the group of Li-Shiuan Peh seeking to improve performance (by using virtual bypassing, a form of routing precomputation) and reducing power consumption (using low-swing signaling).

Actually, if the Ethernet CAT5 does not get to gigabit speeds, then a certification of the cable would fail. That cable is not cat5.

There are two categories of HDMI cables, category-1 and category-2 (for 720p and 1080p, roughly speaking). If a cable is certified for a given cat, it means that it can transmit the corresponding signal with a low (bounded) error rate. Obviously, it is more difficult for a longer cable to pass the cat-2 test. But there are more factors, such as the length, the construction (connectors, etc) and the instalation itself (twists are bad!!).

So no, the fact that you build an Ethernet cable from cat-5 does not mean the final result will be certified for cat-5; similarly, a high-speed HDMI cable should be certified before being sold, and probably the cheapest ones are not. However, the fact that the digital signal transmitted on these cables is uncompressed audio/video makes them much more error-tolerant, so it is likely that a lower-quality cable (i.e., with a higher bit error rate) produces a higher distorsion, but it is not noticed by the end user.

Facebook, on the other side, is mostly public, so the loss of image might also apply.

Oh, wait...

Doesn't it cause blindness? Isn't it the reason for fibers to have power sensing mechanisms that detect broken segments, for safety reasons?

Is this valid only for rooms without windows and without people?

Actually, this is no surprise at all. Maybe most people only focus on the raw speed - i.e., throughput. However, for many applications, the latency - and the lack of sudden latency variations - is more important. These apps are called "inelastic", because they don't tolerate changes in the latency. For example: In a real-time VoIP application, sudden changes in latency make delayed packets useless and the voice gets cut. Yep, you can use a buffer, but that will add an anoying delay in your conversation, so in general the application is highly sensitive to latency changes.

The same happens with games. If you are playing against sb else, your latency can determine if you live or die. AND, the main problem is that the only solution comes from QoS mechanisms that tag, segregate and priorize different flows of traffic. What, I believe, is somehow against net neutrality.

This reminds me of the Spinnaker project, that pretended to simulate a brain (ok, a smaller one, say a fly's brain) in real time. According to their calculations, the processing power of each neuron is very small, so a simple ARM core could handle some 1000 (correct me, this is what I remember) neurons in real time. The complex point was the interconnections between neurons. Obviously, this is much more powerful, despite the 100x slowdown: A much larger brain, and not using specific hardware.