Out of necessity, Edge Machine Learning (ML) moves computing away from the data center to the edges of the network where the data is generated. This means your resource-intensive ML applications must be local applications to function properly.
This raises an important question, which is:
"How do you access the important information on your edge devices?"
You could spend engineering time trying to synchronize the data on your edge devices with your data center, but this is a hard problem to solve, and you can save yourself a lot of time and effort by using Mycelial to move your data where it needs to go.
Mycelial offers you an easy solution to your data movement needs.
Mycelial is an open-source software solution that moves data from sources to destinations.
For example, consider an Edge Machine Learning application that stores its information in a local SQLite database.
So, how do you get the information off of your edge device and onto a system where the information can be analyzed?
Well, with Mycelial you can declaratively create data pipelines that move your data from a source system like SQLite, to a destination system like Snowflake.
There are two main components in Mycelial: clients and a server.
The client is installed and executed on source and destination computers. These clients will register with the server component, and they will receive data pipeline specifications from the server.
The server offers you a way to set up data pipeline specifications, which move your data from one location to another via the installed clients. You can set up these data pipelines via a drag-and-drop web interface, or you can add these pipelines via api calls.
Once you've installed Mycelial (clients and server) you can easily begin moving your data from source systems to destination systems of your choosing. Currently Mycelial has adapters for SQLite but many other adapters are in the works.
Follow our Tutorial to start using Mycelial from the command line.
Watch this short demo:
API documentation
Mycelial is available under the Apache 2 license.