From mboxrd@z Thu Jan 1 00:00:00 1970 X-Spam-Checker-Version: SpamAssassin 3.4.4 (2020-01-24) on polar.synack.me X-Spam-Level: X-Spam-Status: No, score=-1.9 required=5.0 tests=BAYES_00 autolearn=ham autolearn_force=no version=3.4.4 X-Google-Thread: 103376,e7151167e0767ecc X-Google-Attributes: gid103376,public X-Google-Language: ENGLISH,ASCII-7-bit Path: g2news1.google.com!postnews2.google.com!not-for-mail From: wojtek@power.com.pl (Wojtek Narczynski) Newsgroups: comp.lang.ada Subject: Re: Feasibility of using Ada in new development Date: 7 Sep 2004 01:40:26 -0700 Organization: http://groups.google.com Message-ID: <5ad0dd8a.0409070040.60684063@posting.google.com> References: <8429999a.0408231027.2850e800@posting.google.com> <5ad0dd8a.0408302222.56282d6f@posting.google.com> <4135498c_1@news.tm.net.my> <5ad0dd8a.0409040738.3fff41b8@posting.google.com> <5ad0dd8a.0409060956.7fe744f6@posting.google.com> NNTP-Posting-Host: 83.27.46.212 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit X-Trace: posting.google.com 1094546427 11609 127.0.0.1 (7 Sep 2004 08:40:27 GMT) X-Complaints-To: groups-abuse@google.com NNTP-Posting-Date: Tue, 7 Sep 2004 08:40:27 +0000 (UTC) Xref: g2news1.google.com comp.lang.ada:3422 Date: 2004-09-07T01:40:26-07:00 List-Id: Pascal, > Ok, yet you can build such library without modifying the GNAT runtime. There are basically two ways to approach 100K concurrent connections problem, as described in http://citeseer.ist.psu.edu/681845.html: threads and events (there are variants, but those are the two opposite poles). The authors conclude that threads are way easier to program, which my personal experience confirms. With library code the thread (=task) approach is not available. You can't write a library that will allow you to create 100K tasks, nor aid the scheduler with IO descriptor availability awareness. >>> Maybe starting with AWS is a good idea... >> >> Can it scale to > 100K socket connections? > > No, I said that AWS could be a good starting point. polyorb might be a better one, because it already has event demultiplexing code. Regards, Wojtek