<?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 TCP Offloading Engine (TOE) question. in BladeSystem - General</title>
    <link>https://community.hpe.com/t5/bladesystem-general/tcp-offloading-engine-toe-question/m-p/5705057#M28100</link>
    <description>&lt;P&gt;&lt;STRONG&gt;Ed was looking for some semantic clarification:&lt;/STRONG&gt;&lt;/P&gt;&lt;P&gt;&amp;nbsp;&lt;/P&gt;&lt;P&gt;******************&lt;/P&gt;&lt;P&gt;&amp;nbsp;&lt;/P&gt;&lt;P&gt;Hello All…&lt;/P&gt;&lt;P&gt;This may be a matter of semantics… but is there a difference between TCP Offloading Engine (TOE) and this:&lt;/P&gt;&lt;P&gt;&amp;nbsp;&lt;/P&gt;&lt;P&gt;TCP/IP Stateless Offloading&lt;/P&gt;&lt;P&gt;For overall improved system response, the NC375T supports standard TCP/IP offloading techniques including:&lt;/P&gt;&lt;UL&gt;&lt;LI&gt;TCP/IP checksum offload (TCO) moves the TCP and IP checksum offloading from the CPU to the network adapter.&lt;/LI&gt;&lt;LI&gt;Large send offload (LSO) or TCP segmentation offload (TSO) allows the TCP segmentation to be handled by the adapter rather than the CPU.&lt;/LI&gt;&lt;/UL&gt;&lt;P&gt;&amp;nbsp;&lt;/P&gt;&lt;P&gt;&amp;nbsp;&lt;/P&gt;&lt;P&gt;This is for the NC375i and NC375T.&lt;/P&gt;&lt;P&gt;Partner is saying TOE and TCP/IP Stateless Offloading are not the same.&lt;/P&gt;&lt;P&gt;&amp;nbsp;&lt;/P&gt;&lt;P&gt;**************&lt;/P&gt;&lt;P&gt;&amp;nbsp;&lt;/P&gt;&lt;P&gt;&lt;STRONG&gt;Input from Vincent:&lt;/STRONG&gt;&lt;/P&gt;&lt;P&gt;I would agree with the partner, although it is about semantics…&lt;/P&gt;&lt;P&gt;Stateless offloading refers to offloading parts of the TCP protocol processing that do not require maintaining a table of connections (the “state”) in the NIC. It’s easy to calculate the checksum of a packet going through without “knowing” anything about the connection it belongs to.&lt;/P&gt;&lt;P&gt;&amp;nbsp;&lt;/P&gt;&lt;P&gt;TOE without any qualifier could refer to a full TCP/IP stack running on the NIC, including connection establishment/tear down.&lt;/P&gt;&lt;P&gt;&amp;nbsp;&lt;/P&gt;&lt;P&gt;&lt;STRONG&gt;And from Rick:&lt;/STRONG&gt;&lt;/P&gt;&lt;P&gt;I disagree that it is about semantics.&amp;nbsp; TOE is and should always be full offload of the TCP processing - including the "state" of the TCP connection.&amp;nbsp; Stateless offloads are not TOE.&lt;/P&gt;&lt;P&gt;&amp;nbsp;&lt;/P&gt;&lt;P&gt;&lt;STRONG&gt;Reply from Ed:&lt;/STRONG&gt;&lt;/P&gt;&lt;P&gt;Good to know...&lt;/P&gt;&lt;P&gt;Since the QuickSpecs for the NC375 don't specifically call out TOE I would assume it does not support it.&lt;/P&gt;&lt;P&gt;&amp;nbsp;&lt;/P&gt;&lt;P&gt;&amp;nbsp;&lt;/P&gt;&lt;P&gt;&lt;STRONG&gt;And a reply from Rick:&lt;/STRONG&gt;&lt;/P&gt;&lt;P&gt;Probably a good assumption.&amp;nbsp; Further supported by:&lt;/P&gt;&lt;P&gt;&amp;nbsp;&lt;/P&gt;&lt;P&gt;&lt;A href="http://h18004.www1.hp.com/products/servers/networking/nc375t/index.html" target="_blank"&gt;http://h18004.www1.hp.com/products/servers/networking/nc375t/index.html&lt;/A&gt;&lt;/P&gt;&lt;P&gt;&amp;nbsp;&lt;/P&gt;&lt;P&gt;which explicitly says "No" for TOE but does mention LSO (aka TSO) which is a stateless offload.&amp;nbsp; With CKO supported by implication since that is a pre-requisite for TSO/LSO.&lt;/P&gt;&lt;P&gt;&amp;nbsp;&lt;/P&gt;&lt;P&gt;*****************&lt;/P&gt;&lt;P&gt;&amp;nbsp;&lt;/P&gt;&lt;P&gt;&lt;STRONG&gt;Any othe input on this subject?&lt;/STRONG&gt;&lt;/P&gt;</description>
    <pubDate>Thu, 28 Jun 2012 14:21:13 GMT</pubDate>
    <dc:creator>chuckk281</dc:creator>
    <dc:date>2012-06-28T14:21:13Z</dc:date>
    <item>
      <title>TCP Offloading Engine (TOE) question.</title>
      <link>https://community.hpe.com/t5/bladesystem-general/tcp-offloading-engine-toe-question/m-p/5705057#M28100</link>
      <description>&lt;P&gt;&lt;STRONG&gt;Ed was looking for some semantic clarification:&lt;/STRONG&gt;&lt;/P&gt;&lt;P&gt;&amp;nbsp;&lt;/P&gt;&lt;P&gt;******************&lt;/P&gt;&lt;P&gt;&amp;nbsp;&lt;/P&gt;&lt;P&gt;Hello All…&lt;/P&gt;&lt;P&gt;This may be a matter of semantics… but is there a difference between TCP Offloading Engine (TOE) and this:&lt;/P&gt;&lt;P&gt;&amp;nbsp;&lt;/P&gt;&lt;P&gt;TCP/IP Stateless Offloading&lt;/P&gt;&lt;P&gt;For overall improved system response, the NC375T supports standard TCP/IP offloading techniques including:&lt;/P&gt;&lt;UL&gt;&lt;LI&gt;TCP/IP checksum offload (TCO) moves the TCP and IP checksum offloading from the CPU to the network adapter.&lt;/LI&gt;&lt;LI&gt;Large send offload (LSO) or TCP segmentation offload (TSO) allows the TCP segmentation to be handled by the adapter rather than the CPU.&lt;/LI&gt;&lt;/UL&gt;&lt;P&gt;&amp;nbsp;&lt;/P&gt;&lt;P&gt;&amp;nbsp;&lt;/P&gt;&lt;P&gt;This is for the NC375i and NC375T.&lt;/P&gt;&lt;P&gt;Partner is saying TOE and TCP/IP Stateless Offloading are not the same.&lt;/P&gt;&lt;P&gt;&amp;nbsp;&lt;/P&gt;&lt;P&gt;**************&lt;/P&gt;&lt;P&gt;&amp;nbsp;&lt;/P&gt;&lt;P&gt;&lt;STRONG&gt;Input from Vincent:&lt;/STRONG&gt;&lt;/P&gt;&lt;P&gt;I would agree with the partner, although it is about semantics…&lt;/P&gt;&lt;P&gt;Stateless offloading refers to offloading parts of the TCP protocol processing that do not require maintaining a table of connections (the “state”) in the NIC. It’s easy to calculate the checksum of a packet going through without “knowing” anything about the connection it belongs to.&lt;/P&gt;&lt;P&gt;&amp;nbsp;&lt;/P&gt;&lt;P&gt;TOE without any qualifier could refer to a full TCP/IP stack running on the NIC, including connection establishment/tear down.&lt;/P&gt;&lt;P&gt;&amp;nbsp;&lt;/P&gt;&lt;P&gt;&lt;STRONG&gt;And from Rick:&lt;/STRONG&gt;&lt;/P&gt;&lt;P&gt;I disagree that it is about semantics.&amp;nbsp; TOE is and should always be full offload of the TCP processing - including the "state" of the TCP connection.&amp;nbsp; Stateless offloads are not TOE.&lt;/P&gt;&lt;P&gt;&amp;nbsp;&lt;/P&gt;&lt;P&gt;&lt;STRONG&gt;Reply from Ed:&lt;/STRONG&gt;&lt;/P&gt;&lt;P&gt;Good to know...&lt;/P&gt;&lt;P&gt;Since the QuickSpecs for the NC375 don't specifically call out TOE I would assume it does not support it.&lt;/P&gt;&lt;P&gt;&amp;nbsp;&lt;/P&gt;&lt;P&gt;&amp;nbsp;&lt;/P&gt;&lt;P&gt;&lt;STRONG&gt;And a reply from Rick:&lt;/STRONG&gt;&lt;/P&gt;&lt;P&gt;Probably a good assumption.&amp;nbsp; Further supported by:&lt;/P&gt;&lt;P&gt;&amp;nbsp;&lt;/P&gt;&lt;P&gt;&lt;A href="http://h18004.www1.hp.com/products/servers/networking/nc375t/index.html" target="_blank"&gt;http://h18004.www1.hp.com/products/servers/networking/nc375t/index.html&lt;/A&gt;&lt;/P&gt;&lt;P&gt;&amp;nbsp;&lt;/P&gt;&lt;P&gt;which explicitly says "No" for TOE but does mention LSO (aka TSO) which is a stateless offload.&amp;nbsp; With CKO supported by implication since that is a pre-requisite for TSO/LSO.&lt;/P&gt;&lt;P&gt;&amp;nbsp;&lt;/P&gt;&lt;P&gt;*****************&lt;/P&gt;&lt;P&gt;&amp;nbsp;&lt;/P&gt;&lt;P&gt;&lt;STRONG&gt;Any othe input on this subject?&lt;/STRONG&gt;&lt;/P&gt;</description>
      <pubDate>Thu, 28 Jun 2012 14:21:13 GMT</pubDate>
      <guid>https://community.hpe.com/t5/bladesystem-general/tcp-offloading-engine-toe-question/m-p/5705057#M28100</guid>
      <dc:creator>chuckk281</dc:creator>
      <dc:date>2012-06-28T14:21:13Z</dc:date>
    </item>
    <item>
      <title>Re: TCP Offloading Engine (TOE) question.</title>
      <link>https://community.hpe.com/t5/bladesystem-general/tcp-offloading-engine-toe-question/m-p/5957617#M28101</link>
      <description>&lt;P&gt;Does anyone have any benchmark reports or whitepapers which shows the performance increase with/without TOE. I am looking to&amp;nbsp;installing a NC552SFP&amp;nbsp;10GB NICs on a DL580G7. Any help is appreciated&lt;/P&gt;&lt;BLOCKQUOTE&gt;&lt;HR /&gt;&lt;a href="https://community.hpe.com/t5/user/viewprofilepage/user-id/3193"&gt;@chuckk281&lt;/a&gt; wrote:&lt;BR /&gt;&lt;P&gt;&lt;STRONG&gt;Ed was looking for some semantic clarification:&lt;/STRONG&gt;&lt;/P&gt;&lt;P&gt;&amp;nbsp;&lt;/P&gt;&lt;P&gt;******************&lt;/P&gt;&lt;P&gt;&amp;nbsp;&lt;/P&gt;&lt;P&gt;Hello All…&lt;/P&gt;&lt;P&gt;This may be a matter of semantics… but is there a difference between TCP Offloading Engine (TOE) and this:&lt;/P&gt;&lt;P&gt;&amp;nbsp;&lt;/P&gt;&lt;P&gt;TCP/IP Stateless Offloading&lt;/P&gt;&lt;P&gt;For overall improved system response, the NC375T supports standard TCP/IP offloading techniques including:&lt;/P&gt;&lt;UL&gt;&lt;LI&gt;TCP/IP checksum offload (TCO) moves the TCP and IP checksum offloading from the CPU to the network adapter.&lt;/LI&gt;&lt;LI&gt;Large send offload (LSO) or TCP segmentation offload (TSO) allows the TCP segmentation to be handled by the adapter rather than the CPU.&lt;/LI&gt;&lt;/UL&gt;&lt;P&gt;&amp;nbsp;&lt;/P&gt;&lt;P&gt;&amp;nbsp;&lt;/P&gt;&lt;P&gt;This is for the NC375i and NC375T.&lt;/P&gt;&lt;P&gt;Partner is saying TOE and TCP/IP Stateless Offloading are not the same.&lt;/P&gt;&lt;P&gt;&amp;nbsp;&lt;/P&gt;&lt;P&gt;**************&lt;/P&gt;&lt;P&gt;&amp;nbsp;&lt;/P&gt;&lt;P&gt;&lt;STRONG&gt;Input from Vincent:&lt;/STRONG&gt;&lt;/P&gt;&lt;P&gt;I would agree with the partner, although it is about semantics…&lt;/P&gt;&lt;P&gt;Stateless offloading refers to offloading parts of the TCP protocol processing that do not require maintaining a table of connections (the “state”) in the NIC. It’s easy to calculate the checksum of a packet going through without “knowing” anything about the connection it belongs to.&lt;/P&gt;&lt;P&gt;&amp;nbsp;&lt;/P&gt;&lt;P&gt;TOE without any qualifier could refer to a full TCP/IP stack running on the NIC, including connection establishment/tear down.&lt;/P&gt;&lt;P&gt;&amp;nbsp;&lt;/P&gt;&lt;P&gt;&lt;STRONG&gt;And from Rick:&lt;/STRONG&gt;&lt;/P&gt;&lt;P&gt;I disagree that it is about semantics.&amp;nbsp; TOE is and should always be full offload of the TCP processing - including the "state" of the TCP connection.&amp;nbsp; Stateless offloads are not TOE.&lt;/P&gt;&lt;P&gt;&amp;nbsp;&lt;/P&gt;&lt;P&gt;&lt;STRONG&gt;Reply from Ed:&lt;/STRONG&gt;&lt;/P&gt;&lt;P&gt;Good to know...&lt;/P&gt;&lt;P&gt;Since the QuickSpecs for the NC375 don't specifically call out TOE I would assume it does not support it.&lt;/P&gt;&lt;P&gt;&amp;nbsp;&lt;/P&gt;&lt;P&gt;&amp;nbsp;&lt;/P&gt;&lt;P&gt;&lt;STRONG&gt;And a reply from Rick:&lt;/STRONG&gt;&lt;/P&gt;&lt;P&gt;Probably a good assumption.&amp;nbsp; Further supported by:&lt;/P&gt;&lt;P&gt;&amp;nbsp;&lt;/P&gt;&lt;P&gt;&lt;A target="_blank" href="http://h18004.www1.hp.com/products/servers/networking/nc375t/index.html"&gt;http://h18004.www1.hp.com/products/servers/networking/nc375t/index.html&lt;/A&gt;&lt;/P&gt;&lt;P&gt;&amp;nbsp;&lt;/P&gt;&lt;P&gt;which explicitly says "No" for TOE but does mention LSO (aka TSO) which is a stateless offload.&amp;nbsp; With CKO supported by implication since that is a pre-requisite for TSO/LSO.&lt;/P&gt;&lt;P&gt;&amp;nbsp;&lt;/P&gt;&lt;P&gt;*****************&lt;/P&gt;&lt;P&gt;&amp;nbsp;&lt;/P&gt;&lt;P&gt;&lt;STRONG&gt;Any othe input on this subject?&lt;/STRONG&gt;&lt;/P&gt;&lt;HR /&gt;&lt;/BLOCKQUOTE&gt;&lt;P&gt;&amp;nbsp;&lt;/P&gt;</description>
      <pubDate>Thu, 07 Feb 2013 10:49:06 GMT</pubDate>
      <guid>https://community.hpe.com/t5/bladesystem-general/tcp-offloading-engine-toe-question/m-p/5957617#M28101</guid>
      <dc:creator>bhf8888</dc:creator>
      <dc:date>2013-02-07T10:49:06Z</dc:date>
    </item>
  </channel>
</rss>

