Community-Lab introduction

Differences From Artifact [09af443dbe]:

• File script.txt — part of check-in [d6b27da7d8] at 2012-09-19 10:36:30 on branch trunk — Restructure experiment examples to clarify possibilities and usage. (user: ivan, size: 12575) [annotate] [blame] [check-ins using]

To Artifact [d327a8d36a]:

• File script.txt — part of check-in [1e2a1f4994] at 2012-09-19 11:02:39 on branch trunk — Mention link info and management data collection for experiments and external services. (user: ivan, size: 12748) [annotate] [blame] [check-ins using]

   147    147   - Routing: an isolated interface using a VLAN on top of a direct interface.
   148    148     All traffic is allowed, but it can only reach other slivers of the same
   149    149     slice with isolated interfaces on the same physical link.
   150    150   - Low-level testing (not implemented): the sliver is given raw access to the
   151    151     interface.  For privacy, isolation and stability reasons this should only be
   152    152     allowed in exceptional occasions.
   153    153   
          154  +Besides low level access, RDs also offer link quality and bandwidth usage
          155  +measurements for all their interfaces through DLEP (available soon).
          156  +
          157  +Finally, the server and nodes publish management information through an API
          158  +that can be used to study the testbed itself or to implement external services
          159  +(like node monitoring and selection).
          160  +
   154    161   ** An example experiment
   155    162   # Event diagram, hover over components explained.
   156    163   To show how the testbed works: two slivers, one of them (source sliver) pings
   157    164   the other one (target sliver).
   158    165   
   159    166   1. The researcher first contacts the server and creates a slice description
   160    167      which specifies a template for slivers (e.g. Debian Squeeze i386).
................................................................................
   177    184      programs query sliver properties to decide their behaviour.
   178    185   8. Researchers interact straight with containers if needed (e.g. via SSH) and
   179    186      collect results from them.
   180    187   9. When finished, the researcher tells the server to deactivate and
   181    188      deinstantiate the slice.
   182    189   10. Nodes get the instructions and they stop and remove containers.
   183    190   
   184         -At all times there can be external services interacting with researchers,
   185         -server, nodes and slivers, e.g. to help choosing nodes, monitor nodes or
   186         -collect results.
   187         -
   188    191   * Community-Lab integration in existing community networks
   189    192   # CN diagram (buildings and cloud).
   190    193   A typical CN looks like this, with most nodes linked using cheap and
   191    194   ubiquitous WiFi technology (and less frequently Ethernet, optical fiber or
   192    195   others).  The CONFINE project follows three strategies taking into account
   193    196   that CNs are production networks with distributed ownership:
   194    197