Versions Compared

Key

  • This line was added.
  • This line was removed.
  • Formatting was changed.

...

The server design uses a combination of public-key encryption and symmetric-key encryption to accomplish these goals. Symmetric-key encryption is used to encrypt all private information of the user: buddies, goals, devices, etc. Public-key encryption is used to implement a secure messaging system to deliver messages from buddies and messages from the web traffic classification. Every user has two "message boxes", comparable to a classic mailbox with lock:

Image RemovedImage Added

Everyone can drop a message in the box, but you need the key to take a message out of it. Every user has two of these. One is publicly linked to the user account, so it's technically possible to send a message to a user if their identity is known. The other message box is linked indirectly and used for messages related to web traffic. When a user is provisioned, a VPN account is created for them. The user name of this VPN account is a UUID (named accessor ID) that cannot be linked to a user, Through their private key, the user knows what their accessor ID. but without that private key, that link cannot be established. This can be depicted as follows:

...

The server maintains a little bit of unencrypted user information: their name, mobile number and a reference to their public message box. The other information is encrypted with the key that is stored on in the secure storage of the device (e.g. Android non-exported content provider). This encrypted data includes the buddy relationships, the defined goals and two private keys to access the two message boxes. Writing into the message boxes can be done through the public key (not depicted, to prevent confusion). For reading, the private key is required, which is part of the encrypted data of the user. The message box for direct messages is linked from the public data of the user, so given a user, it is possible to find the direct message box. The sole purpose of the direct message box is to be able to send an buddy connect request to a user.

The anonymous message box is linked from the private data of the user, so only the user "knows" their anonymous message box. Besides this, the anonymous message box is also linked to the accessor objectso-called user-anonymized entity, which carries the same ID as the VPN user name. Thus the analysis engine (see below) knows where to post messages when a goal conflict is detected for a given VPN account.  Once the buddy relationship is established, all inter-buddy messages go through the anonymous message box.

See Encryption approach for a more indepth description.

Layered architecture

The Yona server has a simple layered architecture:

...

  • Orange: encrypted with the Yona password stored on the device
  • Magenta: encrypted through public/private key encryption
  • Black: not encrypted. Grey: used to denote conceptual relationships that are in directly programmed this wayThis holds anonymous information that cannot be related to identified users, but users can find the data that belongs to them, by decrypting the associations. The exception to this is the User entity. That is not encrypted  but contains identifieable information: the user name and the e-mail address. This is not encrypted because users need to be able to send a buddy connect request to other users. This information is not exposed through the web APIs. If the system gets hacked, an intruder would be able to fetch a list of names and phone numbers. This is not considered sensitive, as that is also available from public directories. 

User with buddies and goals

...

Various entities carry type stamps. A goal for instance has a creation time and an optional end time, an activity has a start and end time, etc. The convention is that all times are in UTC, except for activities. The design aims to support users that travel through different time zones. To enable that, the entities IntervalActivity and Activity carry a time zone. The start and end of day or week is thus time zone specific.

Cluster architecture

The cluster architecture of the system (both web traffic classification server and Yona server) is covered on a separate page.

Yona consists of 4 different services with different purposes:

  • App service – This is the sole point of contact for the app, providing all backing functions for it.
  • Analysis service – This service analyses the activities of the user and depending on the activity and the goals of the user creates entries in the activity log and/or sends messages to the user and their buddies. These activities are passed on to this service by either the Perl script on the Smoothwall server (in case of network activities) or the app service (in case of app activities)
  • Batch service – Runs scheduled tasks and longer running tasks. Longer running tasks can be triggered by the app service and the admin service.
  • Admin service – This is the point of contact for administrative access to Yona. This service exposes a rudimentary web UI.

The relationships between the services can be depicted as follows:

Image Added

All Yona services use the same database schema and they share a cache, synchronized through Hazelcast. Communication between the services (e.g. from the app service to the analysis service or the batch service) is always through HTTP.

To prevent concurrency issues, the requests of a single user cannot be spread over multiple instances of the analysis service. In case of a multinode cluster, we will need to have a load balancer in front of all services (to take care of load balancing and fail over) and in case of the analysis service, sticky sessions are needed. This is visualized in the below diagram:

Image Added

Note that the admin service talks to the batch service, so that should also be linked to the middle load balancer, but that's omitted from the diagram to prevent too much clutter.