Skip to content

Configuring P2P

This guide explains peer-to-peer (P2P) network mode for Algorand nodes and how to configure your node to enable it. Peer-to-peer networking lets Algorand nodes discover and connect to permissionless Repeaters on a global mesh. Validators make outbound connections to Repeaters, which then propagate blocks, transactions, and participation messages across many independent paths, improving resilience and censorship resistance.

Up until now, Algorand operated only within a more traditional network architecture whereby regular nodes connected to a fixed set of permissioned repeaters that handled message distribution. This created a hierarchy: clients depended on specific servers, and if those servers failed or became congested, the connected nodes lost their pathway to the network. It was like having a few major post offices that all mail must pass through first—efficient when working well, but vulnerable to bottlenecks and single points of failure.

P2P networking, however, adds a mesh layer alongside this existing infrastructure. On Algorand, nodes can now maintain connections to both the established repeater backbone and to community-run Repeaters that they discover dynamically. When a node receives a new transaction or block, it propagates through both the traditional network and the diverse P2P mesh channels, creating redundant paths for the flow of data. This hybrid approach (recommended for most operators) combines the stability of trusted permissioned Repeaters with the resilience and decentralization benefits of permissionless Repeaters. If parts of either system slow down or become unavailable, messages still flow through alternative routes, making the overall network more robust.

  • Greater decentralization: No single infrastructure provider can throttle traffic.
  • Censorship resistance: Transactions and votes propagate through many independent paths.
  • Higher resilience: Resilience is strengthened through path diversity. If a non-P2P Repeater is degraded, P2P Repeaters will still carry traffic, and vice versa.
  • Lower barriers to entry: Anyone with sufficient resources can contribute propagation power.

There are three different P2P configuration modes:

P2P networking is disabled and your node will only connect to Permissioned Repeaters.

Choose this mode if you have strict bandwidth constraints.

Assuming NodeKit is already installed, run the following command:

./bin/nodekit configure algod --hybrid=false

P2P networking is enabled and your node will connect to both Permissionless Repeaters and Permissioned Repeaters.

Choose this mode if you have flexible bandwidth constraints. This mode will also result in the greatest resiliency by connecting to both P2P and non-P2P networks.

Assuming NodeKit is already installed, run the following command:

./bin/nodekit configure algod --hybrid=true

P2P networking is enabled and your node will connect to only Permissionless Repeaters and validators.

Choose this mode if you have very high bandwidth constraints and are able to closely monitor your usage.

NodeKit does not currently support setting P2P to ON (non-Hybrid).

By default, enabling Algorand P2P will only make outbound connections from your node to Repeaters on the mesh. The Repeaters you connect to will see your public IP. Previously, only the Permissioned Repeaters saw it. You can allow inbound connections if you choose, but it is not required. This change improves openness but may not suit everyone running a node from home.

  • Who sees your IP: the specific Repeaters your node connects to in Hybrid or On. In Off, only Permissioned Repeaters can.
  • Inbound connections: off by default; enable only if you want to accept peers.
  • Bandwidth: P2P can increase data usage. Monitor after enabling.
  • Ways to limit exposure: run your node in the cloud or throughx a trusted provider if you prefer not to expose a residential IP.