The author completely misses the point of SPDY - latency reduction by significant reduction in the number of network roundtrips due to slow start, etc. It makes me wonder, how can Varnish claim to be a web accelerator if the author doesn't understand latency?
The author also seems to prioritize his own convenience of development over end user experience (the prime benefit of performance).
Google hasn't had any trouble implementing SPDY proxies. SPDY is actually an opportunity for a proxy like Varnish because anyone could get the performance gain of SPDY just by plugging a SPDY-to-HTTP1 proxy in between.
The author also seems to prioritize his own convenience of development over end user experience (the prime benefit of performance).
Google hasn't had any trouble implementing SPDY proxies. SPDY is actually an opportunity for a proxy like Varnish because anyone could get the performance gain of SPDY just by plugging a SPDY-to-HTTP1 proxy in between.