Why will @ETSI_STANDARDS attempt at non-IP Networking fail? Let us look why it failed in all previous attempts, because #thestandardspeople tried this before. It used to be called NGN or ATM, Future IP, Network 2030 or similar https://www.etsi.org/newsroom/press-releases/1749-2020-04-etsi-launches-new-group-on-non-ip-networking-addressing-5g-new-services

The first thing that ETSI participants forget is that IP stands for Internetworking Protocol. The essence of what mr. @vgcerf tried to achieve was to get different networks talking together. Everybody can make networking protocols (and does), but interconnection is hard

Interconnection is hard because it requires designers to minimize the parameters that are exchanged between networks. You can't make assumptions on service quality, devices, control. You have to assume that despite your network being brilliant (you designed it), the other sucks

ETSI now lists a bunch of stuff TCP/IP can't do. First problem is already that they call it TCP/IP.. there is more than TCP, which is used when TCP isn't practical. For example UDP which is used for a lot of real time stuff.

Efficient use of capacity: Whose capacity and in what context? In 5G networks users are sharing spectrum, which requires different rules of the road then in a fibre network. 5G layer 2 networking sends a boatload of data that isn't necessary on other networks.

Capacity continued: If you don't need to connect to other networks, then you can get rid of all overhead and just send point to point data, like some academic applications do. Every bit is then content. But when you want to send it across different networks, you need overhead.

Security by design: This is impossible in any internetworking context. In your own network you may exert some control, but not on the network of someone else. You have to assume the other guy is an idiot of bad actor. In addition you must assume that end-points can be compromised

Lower Latency for live media: IP has no problem with low latency. Speed of light is an issue, underlying networks, such as LTE do have problems with latency. 5G was supposed to fix that. Ethernet doesn't have that problem much either. Applications however can be hard.

Energy use: yes, that can be an issue, but the issue isn't IP. IP requires hosts to be awake and active. This means constant energy use. But it is hard to do internetworking if you have to assume the other side will be asleep. The other side doesn't know you wanted to speak.

If it is your own network you can fix all these issues to your own liking, but you can't do internetworking with it. This why 5G doesn't get used outside the mobile network and uses IP for internetworking (even between 5G networks.!)

IP scaled to over 60.000 networks and billions of hosts. ETSI/3GPP never got further than 1000 networks with zero internetworking for data. Any advanced services only work over IP. It has actively fought any initiative to let others interconnect their networks to 2G, 3G or 4G.

IP works because it doesn't require Zoom to have contracts with thousands of (corporate) networks. It can just assume that if the IP-link is good enough, stuff will work. Meanwhile ETSI/3GPP/GSMA can't get LTE-M/NB-IOT roaming working.

So let's just assume that @ETSI_STANDARDS is the European Travel and Sightseeing Institute and wastes a ton of EU and telco money on issues that aren't problems. Let's just assume #thestandardspeople are not interested in solutions, but battling issues they lost decades ago.

@threadreaderapp unroll

Here is a blogpost on the website of @cambwireless further explaining this work and all that is wrong with IP https://www.cambridgewireless.co.uk/news/2020/mar/9/initiative-develop-communications-technology-new-d/