Re: [jetty-users] Fast SSL with jetty.

2021-03-21 Thread Greg Wilkins
Luke, well we'll have to look into a JSSE wrapper for OpenSSL again, although Conscrypt has worked fine for some clients, so still worthwhile pushing them to fix any outstanding issues. Also, our priorities are influenced by our commercial support partners, so just a reminder that if there are

Re: [jetty-users] Fast SSL with jetty.

2021-03-21 Thread Luke B
Hi Greg, Realistically I couldn't do the unix socket thing, I think it wouldn't be considered secure enough since clear text sensitive data would be easily available via that socket. Although even if that is not true, I think our clients would just not be ok with anything other than encryption

Re: [jetty-users] Fast SSL with jetty.

2021-03-21 Thread Greg Wilkins
Also I note that it appears that netty is wrapping OpenSSL as a SslEngine, so we could look at either doing the same or even reusing their wrapper (although it appears to pull in a lot of netty util and handlers). On Fri, 19 Mar 2021 at 12:42, Greg Wilkins wrote: > So unix sockets an option? >

Re: [jetty-users] Fast SSL with jetty.

2021-03-18 Thread Greg Wilkins
So unix sockets an option? On Fri, 19 Mar 2021, 09:21 Luke B, wrote: > Hi, > > From memory the difference in performance is rather large, maybe 10x or > 20x. It really does make a difference to how many requests we can handle. > Conscrypt takes jetty from being severely limited by the speed at

Re: [jetty-users] Fast SSL with jetty.

2021-03-18 Thread Luke B
Hi, >From memory the difference in performance is rather large, maybe 10x or 20x. It really does make a difference to how many requests we can handle. Conscrypt takes jetty from being severely limited by the speed at which it can transfer encrypted data, to encryption adding no meaningful

Re: [jetty-users] Fast SSL with jetty.

2021-03-15 Thread Simone Bordet
Hi, On Mon, Mar 15, 2021 at 12:50 AM Luke B wrote: > > Hi, > > So it seems conscrypt has even more memory leaks: > https://github.com/google/conscrypt/issues/835 > https://github.com/google/conscrypt/issues/984 > > Conscrypt doesn't appear to be sufficiently reliable to be used in production. >

Re: [jetty-users] Fast SSL with jetty.

2021-03-15 Thread Shawn Heisey
On 3/14/2021 6:54 PM, Luke B wrote: Unfortunately my clients want that illusion of safety and it is just easier to give them that rather than argue with them. I really don't care to argue this point. I understand. In the case I dealt with, there are certain companies that you simply do not

Re: [jetty-users] Fast SSL with jetty.

2021-03-15 Thread Greg Wilkins
Luke, That memory leak appears to be on outgoing connection attempts, not incoming ones. So conscript should be fine for server side usage with that... or are you also using the client? As for options to avoid conscript, would offloaded SSL that communicates via a unix socket rather than

Re: [jetty-users] Fast SSL with jetty.

2021-03-14 Thread Luke B
Hi Shawn, Unfortunately my clients want that illusion of safety and it is just easier to give them that rather than argue with them. I really don't care to argue this point. Do you have any ideas for solving the problem at hand? cheers, Luke On Mon, Mar 15, 2021 at 11:42 AM Shawn Heisey

Re: [jetty-users] Fast SSL with jetty.

2021-03-14 Thread Shawn Heisey
On 3/14/2021 5:50 PM, Luke B wrote: Setting up jetty to listen only on localhost without SSL and having an nginx (or other web server) reverse proxy to provide SSL is possible but unlikely something that is acceptable as encryption is required all the way to the java process. In this case a

[jetty-users] Fast SSL with jetty.

2021-03-14 Thread Luke B
Hi, So it seems conscrypt has even more memory leaks: https://github.com/google/conscrypt/issues/835 https://github.com/google/conscrypt/issues/984 Conscrypt doesn't appear to be sufficiently reliable to be used in production. Setting up jetty to listen only on localhost without SSL and having