The private network inside a VICIdial cluster, and what crosses it
What actually travels between the machines in a VICIdial cluster, what never touches the private network, and why its speed matters more than its size.

Sooner or later someone asks what actually travels between the machines in a VICIdial cluster, usually the person who looks after your firewall. It's a good question, because the answer explains where a cluster is fragile.
The traffic splits in two. A small set stays inside, on a private network between the machines, and it's the traffic every call waits on.
Inside: the dialers asking the database
Most of it is the dialers talking to the database. Each dialer asks which leads to dial next and records every call it places. When a customer answers, it waits while VICIdial picks a free agent through the database. That hand-off is a round trip across the private network with a customer on the line.
The volume grows with every call. When we replayed a live floor's real database traffic at rising speeds, the query rate climbed in step with the calls, about 12 queries a second for every call up.
| About this many calls up | Database queries a second |
| 180 | 2,155 |
| 540 | 6,464 |
| 900 | 10,774 |
| 1,260 | 15,083 |
Each of those is a question the dialer waits on before it acts. That's what makes this network different from the internet links around it.
Inside: the standby's live copy
The second stream runs from the database to the standby database. The standby replicates the database as it changes, so it always holds a live copy. If the database machine fails, the standby is promoted to take over, a switch called Failover.
In our layout the standby sits on the same private network, in the same datacenter, as the machine it copies. Why a cluster needs a standby database covers when and how it's promoted.
What never touches it
Everything from outside arrives over the internet and lands on the machine that serves it:
- Your carriers connect to the dialers over a SIP trunk.
- Agents' phones, a Softphone app or a browser phone, connect to a dialer too.
- Agents' browsers load their screen from the database and web server, and check in with it every second.
- Managers reach the admin pages and reports on that same machine.
So the private network carries machine-to-machine traffic only. The outside traffic needs its own ports on the public side, and which ports to open for VICIdial lists them.
flowchart LR
subgraph PN["Private network, one datacenter"]
D1[Dialer 1] -->|queries and hand-offs| DB[(Database + web)]
D2[Dialer 2] -->|queries and hand-offs| DB
DB -->|live copy| SB[(Standby database)]
end
C[Carriers] -->|SIP trunks| D1
C -->|SIP trunks| D2
P[Agent phones] --> D1
W[Agent browsers] -->|screen| DBWhy its speed matters more than its size
Every query is a round trip, and the dialer waits for it. What hurts is Latency, the delay on each trip. In the replay, at the floor's real speed, the database answered 99% of queries in 0.45 ms.
A round trip between Germany and Finland takes over 20 ms, more than 40 times that, and it's enough to stall VICIdial. That's why every machine in a cluster sits in one datacenter; the one-datacenter rule explains it in full.
When a shift goes wrong and nobody can say why, we record the next one second by second: database locks, each dialer's load and every call that waited. The VICIdial cluster guide shows how the network fits into the whole layout.
Frequently asked
- No. Agents reach the cluster over the internet: the screen from the database and web server, the phone from a dialer. The private network joins the machines to each other.
- Not in our layout. It replicates over the same private network, in the same datacenter, so it holds a live copy ready to be promoted.
› Do agents connect through the private network?
› Can the standby database sit somewhere else?
If one server covers your floor, there's no internal network to plan, so start on a plan. When you're ready for several machines, book a sizing call and see today's prices.
Related from VICIfast
- Running VICIdial across several serversA database, dialers and a standby, sized from load we measured on a live floor.
- A whole machine for your dialerEvery core yours, mirrored NVMe, delivered in minutes to days. Priced from live stock.
- Bring your own server - $29/moYour VPS or on-prem hardware; we install and manage VICIdial over SSH.
About VICIfast
We run VICIdial servers for call centres, so you don’t have to. You get the dialer set up, secured and kept running — we handle the server, the updates and the backups. Bring your own phone carrier at no markup, or use VICIfast Voice, our own US calling and numbers, from $0.0085 a minute. From $49 a month per server.
Questions? Call +1 636 556 0022 and talk to someone who runs dialers.
Citing this article
VICIfast Engineering. “The private network inside a VICIdial cluster, and what crosses it”. VICIfast LLC, November 21, 2026. Retrieved from https://vicifast.com/blog/vicidial-cluster-private-network
Have questions?
Related posts

Guides & tutorials
Why every server in a VICIdial cluster must sit in one datacenter
Guides & tutorials
Idle campaigns left active: the hidden database load in VICIdial

Guides & tutorials
What is a VICIdial cluster? The parts in plain words

Guides & tutorials
Reports and tools that slow a live VICIdial floor
You might be interested in
Guides & tutorials
When a CRM polls VICIdial once per agent: the load you don't see
Guides & tutorials
Idle campaigns left active: the hidden database load in VICIdial
Guides & tutorials
Why searching a lead by phone number freezes VICIdial, and the missing index

Guides & tutorials
Keep the call log small: log archiving on a busy VICIdial database
VICIfast newsletter
Liked this? Get the next one in your inbox.
We ship the kind of stuff you just read - concrete, numbers-first, no drip. One email when a new post goes live. Unsubscribe in one click.
Comments
No comments yet - be the first.