Are hubs a natural phenomenon in networks or not?

Recently I finished reading the book “linked” by Albert-Laszlo Barabasi.
Probably many of you know it - I found it a pretty good read.

The author claims that the emergence of hubs is a natural phenomenon
in scale-free networks. In those networks, the connectivity of
nodes follows a power-law distribution. meaning that there are many nodes
with a small number of connections but also a few super-nodes or hubs that
have a very large number of connections.
Many examples are given: social networks, the power grid,
neurons in the brain, interaction of species in an ecosystem
aaaaaannnddd… the internet.

For MaidSAFE, often the comparison with ants is made, which work together
in a decentralized manner. But also there, every colony has a queen right?
And every bee-hive has a queen too. And every swarm has a leader (not 100%
sure, I could be wrong on this one).

So I’m wondering: is the design of the internet with its servers and hubs
really so unnatural?

Your thoughts on this?

I’m not the expert on biology, but I think you are wrong on all three - in that while there is a queen bee, its an anthropomorphic label, and does not reflect the actuality. The queen bee does not direct operations - she is the source of eggs. I think in all three cases there is no individual directing behaviour even in a general sense (“let’s try this way”). This is why someone coined the term “hive mind”.

1 Like

Not trying to answer your question but I found the two videos in this post quite educational about bees and their “democratic/consensus” driven ways.

The queen in an ant colony only lays eggs and has zero control. All ants are female, exception in the breeding season when males appear and fly off with potential new queen ants. No central control there :smiley:

2 Likes

Maybe a decentralized app can be like a hub if it attracts enough users and connects them together.

Hubs are normal to a certain extent but they don’t have to be the land we build on.

Whatever attracts the most attention is sorta like a hub of the first order in my opinion. But I don’t think we need leaders of the hive unless you mean the most trusted? You might have to go into detail.

The only issue I could see we could have is the whole web of trust issue which can be solved by a social network app.

1 Like

Trying to straighten this out. Network is the “queen”. Nodes are the “ants” and the data are the “eggs” lol
Had to share

What is the client? Maybe that’s the hatchling of an egg that is queen?

I agree with the social network. Because being social is something all humans have in common so anything that fosters a basic human need can act as a hub. I like the original idea of things being emergent @luckybit. I often think about how the way our planet revolves around the sun, as well as the moon giving us night and day and the moon the tide like our lungs. From this system arose over time (skipping to us) our cells, which are like little cities of organelles, a step up our organs making up our body. Because the cells emerged from the earths rhythms they need rest, so we need rest and so the cities that WE build wake and sleep and have a very human rhythm that seems itself like a living organism. Or think of atoms and their revolving electrons, the solar system revolving, and a galaxy revolving all of which actually encircle a central point but being that there are many (distributed) that are able to interact what emerges is a higher scale of what’s below. One of the things that attracted me to maidsafe is how @dirvine has made such correlations in the design of the network and how intelligence is somewhat emergent of smaller dumbed down systems. I love this project and how philosophically engaging everything is surrounding it

It’s a super old thread, but anyway.

Hubs are simply an attribute of those networks.

Scale-free itself means that the order of nodes follow a power-law distribution, so you’ll see many nodes with few connections, but you’ll also find plenty of nodes with a lot of connections (i.e. more than what you’d expect from a distribution with a thinner tail.) So, hubs don’t “emerge,” they are just a way to describe what a scale-free network looks like.

As for the emergence. Different networks develop through different processes. If new nodes favor well-connected nodes, a scale-free(ish) structure will emerge.

Random errors don’t usually have widespread effect in scale-free networks, so they seem more resilient than they really are. Hubs, however, are easy to find and attack exactly because of their well-connectedness, and the damage can be disproportionate to the attack (let’s just think about the recent DoS attacks against a major DNS provider.)

The good thing about SAFE is that it’s not designed to exhibit scale-free behavior: nodes belong to small groups, and they connect to each other by simple XOR distance metrics, not popularity. Routes may be a little longer this way (hubs are awesome at reducing hop count, though the complete randomness of connections between SAFE nodes may have a similar effect) but in exchange we won’t have to worry about directed attacks through the “weak” form of Single(ish) Point of Failures that hubs are.


It may be unimportant technicality, but most of the networks that are brought up as examples turn out to be not “really” scale-free; instead, they follow other heavy tailed distributions (e.g. exponential), or the power-law only applies to the tail (above a certain threshold.)

1 Like