repo-ng

In current summit, repo sync will start when repo is set up. In next commit, a sync handle command should be invited to instruct repo to start or stop sync process.
Besides, in next commit, a new database to record the current sync status and sync tree should also be added. Otherwise, if a repo died and restart, it will forget all the others status and propagate its data to the synchronized group.

As pointed out on operators mailing list, the lack of repo sync prevents the replication of NDN testbed certificates and makes UCLA router the single point of failure on the NDN testbed.

What's the plan on completing this task?

20160630 conference call discussed this issue.

The existing design is RepoSync: Combined Action-Based and Data-Based Synchronization Model in Named Data Network which needs to be reviewed.

One concern before any implementation is that, the existing repo-ng implementation is of poor code quality.
We should decide whether to add new feature onto it, or rewrite it from ground up before adding new features.

20160719 call states that rewriting repo-ng would not be helpful because no implementation can do much better under the same constraints.