<?xml version="1.0" standalone="no"?>
                    <!DOCTYPE div SYSTEM "/www/backend/www-xml-443/dtd/caidaML.dtd">
                    <!-- do NOT ERASE the DOCTYPE declaration! --><div>


<tr bgcolor="#f4f4f4">
  <td>
<font face="helvetica,arial" size="2">
<b>URL:</b>
</font>
</td>
  <td>
<font face="helvetica,arial" size="2">
<a href="ftp://ftp.ee.lbl.gov/papers/vp-routing-TON.ps.gz">ftp://ftp.ee.lbl.gov/papers/vp-routing-TON.ps.gz</a><br/>
<a href="http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.43.7351">http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.43.7351</a>
</font>
  </td>
</tr>


<tr bgcolor="#e9e9e9">
  <td>
<font face="helvetica,arial" size="2">
<b>Entry Date:</b>
</font>
</td>
  <td>
<font face="helvetica,arial" size="2">
2003-05-15


</font>
  </td>
</tr>


<tr bgcolor="#f4f4f4">
  <td>
<font face="helvetica,arial" size="2">
<b>Abstract:</b>
</font>
</td>
  <td>
<font face="helvetica,arial" size="2">
The large-scale behavior of routing in the Internet has gone virtually
without any formal study, the exceptions being Chinoy's analysis
of the dynamics of Internet routing information [Ch93], and recent
work, similar in spirit, by Labovitz, Malan and Jahanian [LMJ97].
We report on an analysis of 40,000 end-to-end route measurements
conducted using repeated "traceroutes" between 37 Internet sites.
We analyze the routing behavior for pathological conditions, rout-ing
stability, and routing symmetry. For pathologies, we character-ize
the prevalence of routing loops, erroneous routing, infrastruc-ture
failures, and temporary outages. We find that the likelihood
of encountering a major routing pathology more than doubled be-tween
the end of 1994 and the end of 1995, rising from 1.5% to
3.3%. For routing stability, we define two separate types of stabil-ity,
"prevalence," meaning the overall likelihood that a particular
route is encountered, and "persistence," the likelihood that a route
remains unchanged over a long period of time. We find that In-ternet
paths are heavily dominated by a single prevalent route, but
that the time periods over which routes persist show wide varia-tion,
ranging from seconds up to days. About 2/3's of the Internet
paths had routes persisting for either days or weeks. For routing
symmetry, we look at the likelihood that a path through the Internet
visits at least one different city in the two directions. At the end
of 1995, this was the case half the time, and at least one different
autonomous system was visited 30% of the time.


</font>
  </td>
</tr>


<tr bgcolor="#e9e9e9">
  <td>
<font face="helvetica,arial" size="2">
<b>Datasets:</b>
</font>
</td>
  <td>
<font face="helvetica,arial" size="2">
<ul>
<li>traceroute paths between hosts running the "network probe daemon"
    (NPD); 37 participating hosts in 34 stub networks (most are in U.S.)</li>
<li>exponentially distributed measurement intervals</li>
<li>main dataset (D2) collected Nov 3 to Dec 21, 1995:
    <ul>
      <li>60% with mean inter-measurement interval of 2 hours;
          40% with mean interval of 2.75 days</li>
      <li>usually collected paths in both directions for each
          pairing of hosts</li>
    </ul></li>
</ul>


</font>
  </td>
</tr>


<tr bgcolor="#f4f4f4">
  <td>
<font face="helvetica,arial" size="2">
<b>Results:</b>
</font>
</td>
  <td>
<font face="helvetica,arial" size="2">
<ul>
  <li>routing loops in traceroute paths (that is, observe same sequence
      of routers at least <i>three</i> times):
      <ul>
        <li>very few occurrences, but some lasted more than half a day</li>
        <li>can be geographically and temporally clustered</li>
      </ul></li>
  <li>one instance of "erroneous routing," a circuitous path that is clearly
      wrong</li>
  <li>...</li>
</ul>



</font>
  </td>
</tr>
</div>

