DOI :: Mesh networking with Yggdrasil | UPD 4 Social Net (Chapter 1)

 Mesh networking with Yggdrasil

By Bitcoin Cash read.cash11 min

View Original

A couple of months ago I learned of a new technology called Yggdrasil. This is a so

called overlay network that uses mesh-networking. And I think this is very exciting for the broader Internet world.

To understand what the value of Yggdrasil is, let's take a look at what we use today for networking. We use the Internet Protocol, or IP to route TCP over. Together "TCP/IP". When IP version 4 was introduced in 1981 it had a certain design which is largely still unchanged today. We essentially have roads that underpin the Internet of today that are principally unchanged for 40 years now.

As 1981 was really the beginnings of the wider commercial Internet, we can translate that to roads that are like basic cobblestones, meant for horse and carriage to ride over. Roads not wide enough for two cars to pass each other by. Definitely nothing fancy like traffic lights yet.

In contrast to our analogy, our real roads supporting our modern cars have evolved. Many lanes. Lines and indications. Better surface. Traffic signs and lights, to name a few. The roads also evolved in concert with the cars that drive on them.

On the Internet the opposite happened, near zero changes to the basic road system. Limited number of destinations, really bad security etc. Yet, we have a pretty decent Internet today! What gives?

All innovations that actually did happen on the Internet have been done while the roads (IPv4) stayed basically the same. Instead those changes happened in what moves on top of them. For instance, innovative people have created massive private networks to sidestep some limitations on the main one (using NAT).

This may sound harsh to the engineers back then, but it was a different time. Mobile Internet was simply impossible in the mind of the engineers in 1981. What do you expect, put your computer on a cart and stick the cable in each room as you ride by? Don't be silly.

The result of this is that we have to deal with concepts 4 decades old to do new things. Unchanged since 1981 is that an IP-address is determined by location. If I go from my WIFI and walk out the house, my phone gets a new Internet address. The original assumption was that a machine could not move. The fallout is immediate. Changing IP address means you are a different person to the remote server. Connections will time-out waiting for answers from the previous address, which no longer has anyone there. The app on the phone ends up creating a new connection from a new IP address.

To avoid exactly this problem when driving around while the phone hops from one cell-tower to the other, the mobile phone providers' internal network has been setup with quite some overhead to ensure you have one address. A private address from that phone company, not a real one the wider Internet knows you by, mind you.

So, 40 years of innovations and changes of environment have been forced to happen all without actually changing the basic layer 3; the Internet Protocol. And this is very noticeable in our security (the 3-letter agencies that listen in) and ultimately in our user experience. Not to mention cost due to lots of extra innovations that have been added which need to be maintained.

At one point, in 1995, the "next" version of the Internet Protocols (IP) was published: IPv6. Many items that are wrong with IPv4 have been fixed in that version. Stuff I won't mention here other than the most well known, which is the number of addresses world wide. Suffice to say it fixes a lot of other issues as well.

Yet, after 25 years we can't say that the usage of IPv4 has diminished and that IPv6 has stolen marketshare from it. This begs the questions, why is this and how to improve?

Просмотры: