|
1234567891011121314151617181920212223242526272829303132333435 |
- # Bubble Administrator's Guide
-
- # Types of Nodes
-
- A "node" in the land of Bubble is, in the most general sense, any system running the Bubble API and associated services.
- There are a few types of nodes:
-
- ## Regular Nodes
- A regular node is one that is providing VPN and/or Bubble app processing for an end user and their connected devices.
- Regular nodes are launched by other nodes, called Sage Nodes.
-
- ## Sage Nodes
-
- Sage Nodes provides secure proxy connections to various cloud services. A regular node establishes a secure, bilaterally trusted
- relationship with its Sage Node upon launch. A regular node can have only one Sage Node.
-
- ## Peer Nodes (future functionality)
- Today, a Bubble network always contains a single node. In the future, a Bubble can be a cluster of peer nodes.
- An initial regular node would spawn additional peer nodes in accordance with a cluster policy.
-
- End user devices would choose the closest node based on availability and latency.
-
- Bubble network state is stored in PostgreSQL (persistent data) and Redis (cache data).
-
- In order to support multi-node networks, Bubble will:
- * Add a transparent PostgreSQL pooling technology to provide everywhere-hot-active across all peers. Currently, [Pgpool-II](https://www.pgpool.net/) is looking fairly interesting. Open to other suggestions.
- * Add [Redis cluster](https://redis.io/topics/cluster-tutorial) to keep Redis in sync across peer nodes.
- * Possibly: add [etcd](https://etcd.io/) to share configuration information across nodes.
-
- # Start a new Bubble from scratch
-
- * Set up a Bubble [development environment](dev_setup.md)
- * Run the `LaunchSageTest` to launch a new standalone Sage Node
- * Use the new Sage Node to start a new Bubble
-
|
