# Communication between multiple instances of temporal server **URL:** \ **Category:** Community Support\ **Tags:** general-impl, best-practices\ **Created:** [July 20, 2020, 1:29pm UTC](https://community.temporal.io/t/communication-between-multiple-instances-of-temporal-server/228 "2020-07-20T13:29:11Z")\ **Posts on this page:** 1\ **Showing post:** 14
**Author:** ![samar](https://sea2.discourse-cdn.com/flex016/user_avatar/community.temporal.io/samar/32/12_2.png) [@samar](https://community.temporal.io/u/samar)\ **Post date:** [July 21, 2020, 4:50pm UTC](https://community.temporal.io/t/communication-between-multiple-instances-of-temporal-server/228/14 "2020-07-21T16:50:42Z")
Temporal server consists of following 4 roles and the ports needed for those roles for communication. 1. Frontend: This role is responsible for all in-bound API calls, including cross-dc related API which is invoked by remote cluster. It uses the grpcPort `7233` to host service handler and uses port `6933` for membership related communication with other hosts. 2. History: Internal Temporal role to manage workflow state transitions. It uses grpcPort `7234` to host service handler and uses port `6934` for membership related communication. 3. Matching: Internal Temporal role responsible for hosting TaskQueues for task dispatch. It uses grpcPort `7235` to host service handler and uses port `6935` for membership related communication. 4. Worker: Internal Temporal role which runs background processing for replication queue, kafka visibility processor, and system workflows. It uses grpcPort `7239` to host service handler and uses port `6939` for membership related communication. Looks like the following 4 ports in our Dockerfile are unnecessary and probably leftover from Cadence: ```auto 7933 7934 7935 7939 ``` --- _[View the full topic](https://community.temporal.io/t/communication-between-multiple-instances-of-temporal-server/228)._