To me as a web application developer I don't care if I can squeeze 1 - 2 ms more out of a CentOS web server vs. an Ubuntu web server. I'm more concerned about how much more time it will take the dev/ops team to implement the acceptable repos, re-compile updated modules, find or refactor provisioning code to automate the installation across different server instances across the entire development/production server environment. I have seen it take days of extra time (that's a lot of milliseconds) for a RHEL based web server environment to get provisioned ore re-provisioned, tested, and implemented, when there was a pre-built provisioned instance of Ubuntu that could have been tweaked tested and implemented in hours. A few comments on server installs (bloat). First, All enterprise Linux server installs have a GUI based (bloated) install option and most have a minimal install option -- including Ubuntu. At the Ubuntu install screen hit F4 and choose minimal, and you will get a completely manual install process (with a smaller installer footprint then SuSE). All additional server applications including OpenSSL, OpenSSH, LAMP, Samba, mail, amongst others will only be installed if you select them to be. When finished you will have a headless/CLI only instance of the server, No GUI, DE etc. Why would you want a GUI on your server anyway? to me that's bloat. Second, A bigger question regarding installation - In a commercial setting who uses the installer anymore? Corporate IT based provisioned or pre-provisioned snapshots are the standard. Most if not all server instances are reduced to as few tasks as necessary as a part of the provisioning. In other words there is USUALLY no bloat installed. Use case and provisioning are the new key considerations. With the cloud driving IT, gone are the days when it is productive (in a commercial instance) to simply take a version of RHEL (or insert any single distro here) and use it alone to hammer out any type of server you may need. In other words the cost in human capital to provision a web server (THE USE CASE) becomes the decision factor not the performance of one version of this server vs the other. I'm not suggesting that CentOS or RHEL are not good web servers or that in all cases Ubuntu is better than these. I use CentOS and RHEL everyday as well as Ubuntu. But to me use case is key. In human terms whether a web server can deliver a page 1, 2, or even 10 ms faster is not relevant because to the human being there is no noticeable difference in speed. However if the server is not a web server but mining bitcoin or processing thousands of stock trades per second than 10 ms is huge. Just my thoughts On Wed, Jan 21, 2015 at 10:12 AM, Eric Cope wrote: > > http://www.phoronix.com/scan.php?page=article&item=ubuntu_1404_distros1&num=1 > > On Wed, Jan 21, 2015 at 10:00 AM, Nathan England wrote: > >> >> Sorry to fan the flames but I am curious... We have had some talk lately >> about web servers and ubuntu/centos and various applications needing some >> version or another. >> >> I'm curious, has anyone ever seen a performance comparison between >> similarly configured ubuntu and centos systems? >> >> Should there be a performance difference? I'm sincerely curious here, I >> am interested about vps performance differences as well as physical server >> differences. >> >> Nathan >> --------------------------------------------------- >> PLUG-discuss mailing list - PLUG-discuss@lists.phxlinux.org >> To subscribe, unsubscribe, or to change your mail settings: >> http://lists.phxlinux.org/mailman/listinfo/plug-discuss >> > > > --------------------------------------------------- > PLUG-discuss mailing list - PLUG-discuss@lists.phxlinux.org > To subscribe, unsubscribe, or to change your mail settings: > http://lists.phxlinux.org/mailman/listinfo/plug-discuss > -- James *Linkedin *