|
| 1 | +--- |
| 2 | +title: "Bacalhau" |
| 3 | +description: "A Computer Over Data Framework for IPFS" |
| 4 | +draft: false |
| 5 | +menu: |
| 6 | + curriculum: |
| 7 | + parent: "curriculum-devtools" |
| 8 | +category: lecture |
| 9 | +weight: 80 |
| 10 | +level: |
| 11 | +- shallow |
| 12 | +- deep |
| 13 | +--- |
| 14 | + |
| 15 | +Bacalhau is a framework that allows you to perform distributed computations on IPFS. |
| 16 | +These computations are executed in the form of Docker containers or WASM programs. |
| 17 | + |
| 18 | +The **Bacalhau network is a set of computers (nodes) working ready to perform computations**. |
| 19 | +You can send your computation (Docker container or WASM) to the Bacalhau network and then read the results. |
| 20 | +Every computation has an _input_ and an _output_. |
| 21 | +In Bacalhau, the main idea is that the **input is taken from IPFS, and the output is published to IPFS**. |
| 22 | + |
| 23 | +For example, consider that you have a Python script stored on IPFS, with CID `QmR1...`. |
| 24 | +This script prints numbers from 0 to 100, and you want to execute it on the Bacalhau network. |
| 25 | + |
| 26 | + |
| 27 | + |
| 28 | +First, you provide the CID of the Python script to the Bacalhau network. |
| 29 | +The script is fetched from IPFS, and is executed by one or several Bacalhau nodes. |
| 30 | +After the computation is finished, the results (the numbers printed) are published to IPFS, with a new CID, `QmFH...`. |
| 31 | + |
| 32 | +## Architecture |
| 33 | + |
| 34 | +Every node in the Bacalhau network can act as a _requester node_ or a _compute node_. |
| 35 | +Requester nodes are responsible for receiving the computations (in technical terms, called **jobs**) and delegating them to one or several compute nodes. |
| 36 | +Compute nodes are responsible for performing the actual computation. |
| 37 | + |
| 38 | + |
| 39 | + |
| 40 | +The lifecycle of a job in Bacalhau is a complex task that includes several steps and verifications. |
| 41 | +You can read everything about the Bacalhau architecture in the [official documentation](https://docs.bacalhau.org/about-bacalhau/architecture). |
| 42 | + |
| 43 | +## The Bacalhau CLI |
| 44 | + |
| 45 | +Sending your computations and communicating with the Bacalhau network is pretty easy by using the Bacalhau CLI. |
| 46 | +The `bacalhau` command installed on your computer allows you to send a new job (computation), retrieve its status, or get its results. |
| 47 | + |
| 48 | +```bash |
| 49 | +> bacalhau docker run ubuntu echo Hello World |
| 50 | +``` |
| 51 | + |
| 52 | +The previous command creates a new job in Bacalhau network by using a Docker image. |
| 53 | +Specifically, it runs an Ubuntu container and executes `echo Hello World` inside the container. |
| 54 | + |
| 55 | +You can then verify the status of the job by executing the following command. |
| 56 | + |
| 57 | +```bash |
| 58 | +> bacalhau list --id-filter=${JOB_ID} |
| 59 | +``` |
| 60 | + |
| 61 | +If you are familiar with Docker, you should identify the `docker run` command, which allows you to execute Docker containers. |
| 62 | +The Bacalhau CLI adds several options on top of this command. |
| 63 | + |
| 64 | +```bash |
| 65 | +> bacalhau docker run \ |
| 66 | +-v QmfKJT13h5k1b23ja3ZCVg5nFL9oKz2bVXc8oXgtwiwhjz:/files \ |
| 67 | +ubuntu cat /files/read_csv.py |
| 68 | +``` |
| 69 | + |
| 70 | +In the previous example, the command `cat /files/read_csv.py` is executed inside an Ubuntu container. |
| 71 | +The `-v` (volume) option allows you to mount an IPFS file or folder to a directory inside the Docker container. |
| 72 | +The IPFS folder with CID `QmfKJT13h5k1b23ja3ZCVg5nFL9oKz2bVXc8oXgtwiwhjz` is mounted to the `/files` directory inside the container. Therefore: |
| 73 | +1. An Ubuntu Docker container is created |
| 74 | +2. The `QmfKJT13h5k1b23ja3ZCVg5nFL9oKz2bVXc8oXgtwiwhjz` IPFS folder is mounted to the `/files` directory inside the container. |
| 75 | +3. The file is read by using the `cat` command. |
| 76 | + |
| 77 | +Now that you understand the basics of Bacalhau, check out the great examples in the [official documentation](https://docs.bacalhau.org/examples/). |
0 commit comments