[openssl-users] [EXT] TLS 1.3 performance

[Roelof Du Toit]

Hi Neetish,

I would recommend considering the following in your research:
- The impact of Nagle.  See https://github.com/openssl/openssl/issues/4237
- The impact of the KeyShare calculation on TLS 1.3 session resumption (assuming most deployments will use psk_dhe_ke)
- The impact of post-handshake handshake messages.

Regards
Roelof

From: openssl-users <openssl-users-bounces at openssl.org> on behalf of Neetish Pathak <npathak2 at ncsu.edu>
Reply-To: "openssl-users at openssl.org" <openssl-users at openssl.org>
Date: Thursday, September 14, 2017 at 4:20 PM
To: "openssl-users at openssl.org" <openssl-users at openssl.org>
Subject: [EXT] [openssl-users] TLS 1.3 performance

Hi,

I worked on TLS 1.3 performance bench-marking. After my tests, I found that TLS 1.3 based resumption is not giving us the connection latency benefits when tested in a LAN environment. It is slower than TLS 1.2. When tested on WAN, definitely, TLS 1.3 fares better than TLS 1.2.
I want your suggestion on whether can I work on the following problem for my research project: "Investigating why TLS 1.3 resumption/0-RTT/full-handshake doesn't provide the  desired connection latency benefits in LAN and the ways to make it faster?"
I want suggestions/references on the right research problem. Any ideas will be appreciated.

Thanks in advance

Thanks
Best Regards,
Neetish
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mta.openssl.org/pipermail/openssl-users/attachments/20170914/2478e3de/attachment.html>