Gitiles
Code Review Sign In
code.v1.dodo.cloud / pcloud / 2babef2bf7d37f5b36627e1b5ec144c35254cfde
commit2babef2bf7d37f5b36627e1b5ec144c35254cfde[log] [tgz]
authorgiolekva <giolekva@gmail.com>Wed Mar 25 23:27:29 2020 +0400
committergiolekva <giolekva@gmail.com>Wed Mar 25 23:27:29 2020 +0400
tree22d922f158b94ae0ac4a18addc3df6b27abf5ed0
parentd2a029dfe4713ad6ae7c79dba46444f244ec09a2 [diff]
introduce replica assignment change listener
5 files changed
tree: 22d922f158b94ae0ac4a18addc3df6b27abf5ed0
  1. api/
  2. chunk/
  3. client/
  4. docs/
  5. k8s/
  6. master/
  7. testing/
  8. .gitignore
  9. chunk_server.go
  10. client.go
  11. Dockerfile
  12. master.go
  13. README.md
README.md

PCloud

PCloud is a set of distributed infrustructure tools meant for setting up a personal cloud on commodity hardware such as Raspberry Pi. Goal of the project is to provide:

  • Blob storage: distributed file storage with configurable replication for redundancy
  • Namespace service: built on top of Blob storage to expose it, at least in read only mode, as a mountable file system
  • Metadata service: storing Blob annotations such as objects and their locations detected in the pictures
  • Search infrastructure: to index Blob metadata and provide search API
  • App infrastructure: to easily set up and run third party applications such as custom in picture object detection models

To prove viability of the project first milestone will be to build fully functional image storage and indexing infrustructure. User experience setting it up will look something like:

  • Set up PCloud on 3 or more servers and pair mobile device with it
  • Configure IFTTT (if this then that) like worklfow to automatically:
    • Back up every newly taken picture on PCloud
    • Run face detection app on backed up pictures and store this information in Metadata service
    • Index face annotations and make them searchable

User must be able to configure all of these from previously paired mobile device.