<?xml version="1.0" encoding="UTF-8"?>
<rss xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#" xmlns:taxo="http://purl.org/rss/1.0/modules/taxonomy/" version="2.0">
  <channel>
    <title>topic Re: Performance question? 150 node network with multiple switches... in Operating System - HP-UX</title>
    <link>https://community.hpe.com/t5/operating-system-hp-ux/performance-question-150-node-network-with-multiple-switches/m-p/3025755#M574103</link>
    <description>Thanks Ron.. I will try to give it a shot this week..  Appreciate all of the advice..</description>
    <pubDate>Thu, 17 Jul 2003 14:28:17 GMT</pubDate>
    <dc:creator>Dale Sullens</dc:creator>
    <dc:date>2003-07-17T14:28:17Z</dc:date>
    <item>
      <title>Performance question? 150 node network with multiple switches...</title>
      <link>https://community.hpe.com/t5/operating-system-hp-ux/performance-question-150-node-network-with-multiple-switches/m-p/3025748#M574096</link>
      <description>I have a 150 node network with approximately 50 workstations and 100 Servers, all on 1 floor with no network segmentation.&lt;BR /&gt;&lt;BR /&gt;The configuration (this was not by design) I am working with consists of a primary 80 Port HP procurve switch with 4 smaller switches at locations of high concentration.&lt;BR /&gt;My question is how much of an impact is the overall network performance affected by these "extra" switches?&lt;BR /&gt;&lt;BR /&gt;The building we are in right now only has 1 or 2 network drops per room so I am limited on where I can place my equipment.  Hence I may have 1 network cable from the "main" switch feeding 30 Servers...&lt;BR /&gt;&lt;BR /&gt;I would like to have 1 drop per node to the "main" switch. Without some hard facts it will be difficult to convince management to spring for any cabling changes in the building..&lt;BR /&gt;&lt;BR /&gt;Thanks for any advice..</description>
      <pubDate>Wed, 16 Jul 2003 18:15:59 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-hp-ux/performance-question-150-node-network-with-multiple-switches/m-p/3025748#M574096</guid>
      <dc:creator>Dale Sullens</dc:creator>
      <dc:date>2003-07-16T18:15:59Z</dc:date>
    </item>
    <item>
      <title>Re: Performance question? 150 node network with multiple switches...</title>
      <link>https://community.hpe.com/t5/operating-system-hp-ux/performance-question-150-node-network-with-multiple-switches/m-p/3025749#M574097</link>
      <description>It sort of depends on the size of the links between switches and the amount of traffic.  &lt;BR /&gt;&lt;BR /&gt;Get mrtg and let it monitor your links between switches.  It will show you where the bottlenecks are in pretty graphs which even managers can understand.&lt;BR /&gt;&lt;BR /&gt;people.ee.ethz.ch/~oetiker/webtools/mrtg&lt;BR /&gt;&lt;BR /&gt;It would be interesting to know what level of broadcast traffic you are running into on your network.  If these are all Window boxes it could be a considerable proportion of your available bandwidth.  Reducing the worthless broadcasts might help.  How many of your machines are running netbeui for no reason?&lt;BR /&gt;&lt;BR /&gt;Ron&lt;BR /&gt;&lt;BR /&gt;</description>
      <pubDate>Wed, 16 Jul 2003 18:23:42 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-hp-ux/performance-question-150-node-network-with-multiple-switches/m-p/3025749#M574097</guid>
      <dc:creator>Ron Kinner</dc:creator>
      <dc:date>2003-07-16T18:23:42Z</dc:date>
    </item>
    <item>
      <title>Re: Performance question? 150 node network with multiple switches...</title>
      <link>https://community.hpe.com/t5/operating-system-hp-ux/performance-question-150-node-network-with-multiple-switches/m-p/3025750#M574098</link>
      <description>This kind of setup is surely going to be a drag on performance.&lt;BR /&gt;&lt;BR /&gt;Even a ping request has a lot of workstations to poll to get an answer.&lt;BR /&gt;&lt;BR /&gt;Try pinging the broadcast address and see how long that takes to cycle.&lt;BR /&gt;&lt;BR /&gt;More important, during periods of high network load, bottlenecks can develop.&lt;BR /&gt;&lt;BR /&gt;Toptools, is a good way to monitor the network configuration itself, but it only runs on certain HP hardware.&lt;BR /&gt;&lt;BR /&gt;A traceroute (tracert in windows) command can be used to look for bottlenecks.&lt;BR /&gt;&lt;BR /&gt;Utilization on that server wire may be a problem as well and needs to be looked at.&lt;BR /&gt;&lt;BR /&gt;I'm not familiar with that HP switch, but a modern Cisco Switch to segment the servers and provide a faster Core backbone can really improve performance a lot.&lt;BR /&gt;&lt;BR /&gt;SEP</description>
      <pubDate>Wed, 16 Jul 2003 18:28:22 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-hp-ux/performance-question-150-node-network-with-multiple-switches/m-p/3025750#M574098</guid>
      <dc:creator>Steven E. Protter</dc:creator>
      <dc:date>2003-07-16T18:28:22Z</dc:date>
    </item>
    <item>
      <title>Re: Performance question? 150 node network with multiple switches...</title>
      <link>https://community.hpe.com/t5/operating-system-hp-ux/performance-question-150-node-network-with-multiple-switches/m-p/3025751#M574099</link>
      <description>90% of my systems are Windows based.. The remaining are a mix of Unix boxes..  &lt;BR /&gt;I'll give mrtg a try. &lt;BR /&gt;&lt;BR /&gt;Whats the easiest way to disable Netbeui?&lt;BR /&gt;&lt;BR /&gt;I'm still running a Windows NT4 domain.&lt;BR /&gt;&lt;BR /&gt;&lt;BR /&gt;Thanks</description>
      <pubDate>Wed, 16 Jul 2003 18:41:50 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-hp-ux/performance-question-150-node-network-with-multiple-switches/m-p/3025751#M574099</guid>
      <dc:creator>Dale Sullens</dc:creator>
      <dc:date>2003-07-16T18:41:50Z</dc:date>
    </item>
    <item>
      <title>Re: Performance question? 150 node network with multiple switches...</title>
      <link>https://community.hpe.com/t5/operating-system-hp-ux/performance-question-150-node-network-with-multiple-switches/m-p/3025752#M574100</link>
      <description>Some switches let you block anything but IP otherwise I think it's probably a one machine at a time thing.  You could put a sniffer on it and see if you are getting a lot of netbeui going over the interswitch links.  The switches themselves may be able to give you some info on the amount of broadcast traffic which you can commpare to the total number of packets in or out.  Some of them allow you to limit the percentage of broadcast traffic.&lt;BR /&gt;&lt;BR /&gt;Ron</description>
      <pubDate>Wed, 16 Jul 2003 19:02:11 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-hp-ux/performance-question-150-node-network-with-multiple-switches/m-p/3025752#M574100</guid>
      <dc:creator>Ron Kinner</dc:creator>
      <dc:date>2003-07-16T19:02:11Z</dc:date>
    </item>
    <item>
      <title>Re: Performance question? 150 node network with multiple switches...</title>
      <link>https://community.hpe.com/t5/operating-system-hp-ux/performance-question-150-node-network-with-multiple-switches/m-p/3025753#M574101</link>
      <description>These switches are pretty low end( except for the main HP pro curve) I would appreciate a  recommendation for a network sniffer?&lt;BR /&gt;&lt;BR /&gt;&lt;BR /&gt;Thanks!</description>
      <pubDate>Wed, 16 Jul 2003 19:17:22 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-hp-ux/performance-question-150-node-network-with-multiple-switches/m-p/3025753#M574101</guid>
      <dc:creator>Dale Sullens</dc:creator>
      <dc:date>2003-07-16T19:17:22Z</dc:date>
    </item>
    <item>
      <title>Re: Performance question? 150 node network with multiple switches...</title>
      <link>https://community.hpe.com/t5/operating-system-hp-ux/performance-question-150-node-network-with-multiple-switches/m-p/3025754#M574102</link>
      <description>tcpdump is what all of the UNIX types like.  &lt;A href="http://www.tcpdump.org" target="_blank"&gt;www.tcpdump.org&lt;/A&gt; or get the official HP version form the HP website.&lt;BR /&gt;&lt;BR /&gt;If you are running a Window box you can use either Windump &lt;BR /&gt;&lt;A href="http://windump.polito.it/install/default.htm" target="_blank"&gt;http://windump.polito.it/install/default.htm&lt;/A&gt;&lt;BR /&gt;or Snort (&lt;A href="http://www.snort.org)." target="_blank"&gt;www.snort.org).&lt;/A&gt;  Both require Winpcap to run.  Snort's really an intrusion detector but it can do lots of other things.&lt;BR /&gt;&lt;BR /&gt;In etiher case since you are working with switches you will need to tell the switch to let you monitor the traffic on port x by way of port y then you plug your box into port y.  Port x should be one of your interswitch links.  Probably your core switch would be your best bet.&lt;BR /&gt;&lt;BR /&gt;If you have any trouble getting mrtg to work let me know.  You will need to tell it the ip address of the management interface for each switch as well as the community string.  If your cheap switches are all connected directly to the big switch then you really only have to monitor the big switch.  The big switch may need to know the ip address of the box you are running mrtg on if you have set up security on it.&lt;BR /&gt;&lt;BR /&gt;Ron</description>
      <pubDate>Wed, 16 Jul 2003 19:35:13 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-hp-ux/performance-question-150-node-network-with-multiple-switches/m-p/3025754#M574102</guid>
      <dc:creator>Ron Kinner</dc:creator>
      <dc:date>2003-07-16T19:35:13Z</dc:date>
    </item>
    <item>
      <title>Re: Performance question? 150 node network with multiple switches...</title>
      <link>https://community.hpe.com/t5/operating-system-hp-ux/performance-question-150-node-network-with-multiple-switches/m-p/3025755#M574103</link>
      <description>Thanks Ron.. I will try to give it a shot this week..  Appreciate all of the advice..</description>
      <pubDate>Thu, 17 Jul 2003 14:28:17 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-hp-ux/performance-question-150-node-network-with-multiple-switches/m-p/3025755#M574103</guid>
      <dc:creator>Dale Sullens</dc:creator>
      <dc:date>2003-07-17T14:28:17Z</dc:date>
    </item>
  </channel>
</rss>

