ann on monday, page-129

Riverbeds will do packet level optimisation of encrypted data but typically get extremely poor optimisation rates due to the highly randomised nature of the data (also why packet block dedupe sucks on crypto data also). This is why Steelheads need to MITM to strip SSL at each terminating device to get to the real, compressible content.

Given Wangle presumably don't MITM (would be a massive privacy issue) I'm still unsure what their plan is to optimise encrypted data. With HTTP/2 being TLS only this would be a showstopper issue if they can't. Ergo If they've invented a method to optimise completely randomised (patternless, undeduplicatable) data efficiently, then the market cap should be much higher as it would blow their competition out the window!

We'll be stuck with unencrypted HTTP1.1 for a good while yet though. Enough time to make some good revenue.