CliqSafe – How Do We Do This?

CLIQSAFE: STAGE  1

  1.   Interrogates user email for all URLs
  2.   Packages them up in a small JSON call to CliqSafe Cloud through port 80 (or other port). May or may not require firewall rule changes on client-side.
  3.   Package is pre-parsed in CliqSafe Cloud against Client-generated whitelists, SEVERAL SIMPLE LOGICAL RULES, & Google Safe Browsing List(in order to reduce latency, parsing is done using node.js AWS λ (Lambda) API’s, or Microsoft Azure Cloud Functions.
  4.   Queries optimized on Tez (We think).
  5.   If STAGE 1 tests can make a decision w/ > 90% certainty, then when STAGE 1 returns result to CliqSafe & results are returned to client through port 80 (or other) as an HTML overlay.
  6.   If results are not positive, then suspicious URLs move to STAGE 2

CLIQSAFE: STAGE  2

  1.  Stage 2: The AI

2.  Using vagrant, ~100 minimal Debian Stretch VM’s to be spun up at all times

3.  If URL == suspicious && ! in GSBL; then request URL is routed to wget batch script on a single VM

4.  wget batch script requests actual suspicious URL.

5.  When request is received, rules are applied to returned results of HTTP GET request.

6.  Then, we do the math…  Fast.