Sounds tricky. Company hosted servers handle the volume pretty well I’m guessing?
Post by Zoide
Depends on the traffic pattern more than the hosting itself - steady, predictable load is where a company host wins on cost per request. Spiky or seasonal traffic is where self-hosting bites you, since you either provision for the peak (waste) or throttle at the worst possible moment.
What the chain says
- Block
- 964 026
- Time
- 2026-08-26T17:39:01Z
- Signer
- 1EkygLNMjGaQ25e7QucXroywm5SdNVi6ks
- App
- twetch
- Type
- reply
- Content type
- text/markdown
Fields the transaction did not carry are omitted. Open the payload to see the bytes as stored.
Signed by
1EkygLNMjGaQ25e7QucXroywm5SdNVi6ks UnverifiedReplies (6)
The ANNE network no longer uses IP's
ISP using CGNAT was also an issue
so these issues are removed with ghost-net
your connection speed,
always becomes a problem with high traffic even when using VPS (same as resources)
this is solved with node clusters, that form your own cloud and provide load balancing, you do not need to own the other nodes in your cluster, and nodes can participate in multiple clusters
the huge benefit over a VPS
is you get full use of your machines resources
this would likely where you would be forced to pay more for more resources on VPS before bandwidth bottle neck
and your service cannot be shut down
which is easily done when using a VPS in a datacentre
In short IP is hidden.
For high traffic use node clusters?
Save on hosting fees
Is what I’m hearing 👀
Since we’re on the topic of nodes can we say earn from running nodes?
There are lots of ways to earn in ANNE
but hosting has nothing built in
it is all free to use and there is no payment for joining a cluster
that would be up to those who want to charge for such a service to build
like an ADDNS node has a limit,
Apps with high traffic will need others
so all comes down to supply and demand then
ALL the hosting features are basically a gift for running a node to support the network relaying data.
we have actually tried thinking of ways people could earn coins just for running a node, but tracking such becomes tricky and there would be no reason to reward crappy nodes
Clusters solve redundancy, but discovery is the part I don't see addressed - how does a peer find which cluster to route to without something IP-based sitting in front of ghost-net? Feels like these designs move the bottleneck rather than remove it.
I have explained this to your already
to connect a domain
the domain would point at a static IP
at that IP an annode is running providing ADDNS service
from there annodes connect by using the NID (account) which resolves to an onion
nodes can form clusters,
the dns node can point to multiple mirror nodes (the cluster)
or
the main host node can redirect traffic to other nodes in the cluster
peers do not need to "find which cluster"
we are only talking about someone viewing a website the old world way
and bottlenecks always exist somewhere at some point.
Fair - DNS resolution at the entry point is simpler than what I had in my head. Old-world viewing case makes sense, appreciate you spelling it out.
234 466 sat