Several long-running end-user services run continuously on PlanetLab,
generating over 4 TBytes of live network traffic and contacting over
1 million unique IP addresses every day. Users at PlanetLab sites (and
elsewhere) are encouraged to take advantage of these services.
Content Distribution Networks
|
Service |
MBytes/day |
Destination IPs/day |
| CoralCDN | n/a | n/a |
Distributed Hash Tables
|
Service |
MBytes/day |
Destination IPs/day |
Scalable Large-File Transfer
|
Service |
MBytes/day |
Destination IPs/day |
Robust DNS Resolution
|
Service |
MBytes/day |
Destination IPs/day |
Routing Overlays
|
Service |
MBytes/day |
Destination IPs/day |
Multicast
|
Service |
MBytes/day |
Destination IPs/day |
Email
|
Service |
MBytes/day |
Destination IPs/day |
Publish-Subscribe News Access
|
Service |
MBytes/day |
Destination IPs/day |
Mobile Access
|
Service |
MBytes/day |
Destination IPs/day |
Network Measurement & Observation
Many researchers use PlanetLab as an observation platform from
which they can measure and map the Internet, detect anomalies, and
diagnose failures. Few of the tools developed to support this
activity are available as continuously running PlanetLab services.
The main exception is ScriptRoute, which allows users to execute
measurement scripts from PlanetLab's main vantage points.
Infrastructure Services
PlanetLab service developers commonly take advantage of "helper"
services, commonly called infrastructure services, to deploy, monitor,
and maintain their services and experiments. New PlanetLab users are
strongly encouraged to leverage these services, rather than reinvent
their own.
Brokerage Services
These services can be used to acquire additional resources, for
example, to run an experiment or demo during periods of heavy
load.
Environment Services
These services are used to distribute software packages, resolve
dependencies, and remotely control a slice. Users should be aware that
there are many tools available. We call these two out because the combination seems
to be getting the most traction in the community. It would be a great
service to the PlanetLab community for someone to write a "best
practices" document that helps new users figure out what combination
of deployment and management tools to use for their slice.
Monitoring & Discovery Services
These are related services that can either be used to monitor the
behavior running nodes or slices, or to help users discover what
resources are available on what nodes.
