The Kentucky Aggregate
An initial version of the Kentucky aggregate is up and running and was demonstrated working with the Utah clearinghouse and the Utah aggregate at the GEC 4 meeting. The Kentucky aggregate consists of 79 Dell PCs. Of the 79 machines, 47 are Dell Optiplex GX240 available for general use, and 24 are Dell Inspiron 530 machines which are currently reserved for work on the InstrumentationTools Project. An additional 8 Dell Poweredge 1900 machines were recently added to the list of available machines and are being used for student class projects. All machines have 4 experimental network interfaces that can be configured via the Emulab web interface to produce the desired topology, bandwidth, delay, etc.
Accessing The Kentucky Aggregate
The Kentucky ProtoGENI aggregate can be accessed by obtaining an account at one of the ProtoGENI aggregates. The preferred way to gain access to ProtoGENI is to obtain an account on the Utah aggregate at http://emulab.net. Access to the Utah aggregate will also enable access to the Kentucky aggregate. However, users who plan to primarily use resources on the Kentucky aggregate may apply for an account at the Kentucky site http://www.uky.emulab.net directly.
Once this account is created, a user who wishes to access ProtoGENI resources should create a personal key that will be used to authenticate them to ProtoGENI. Key creation is described under "Getting Ready" on the ProtoGENI wiki (http://www.protogeni.net/trac/protogeni/wiki/Tutorial). In short, the new user must login to their Emulab account and following the "My Emulab->Profile->Generate SSL Cert" links to create a unique certificate. After giving a secure pass-phrase, your certificates will be generated and saved in $HOME/.ssl/encrypted.pem. These credentials can be used to access any of the ProtoGENI aggregates (including the Kentucky aggregate).
Obtaining ProtoGENI resources is easy using FLACK. Once the experiment has been created, adding INSTOOLS is as simple as pressing a button. see Tutorial: Using the Flack and Instrumentation Portal.
In the early development stages, the interface used to obtain resources was rather primitive. To obtain resources users had to write a python script that makes calls to the ProtoGENI API. A few test scripts are available as examples. The ProtoGENI API is described at http://www.protogeni.net/trac/protogeni/wiki/ComponentManagerAPI, and can be used to create slices, slivers, and boot those slivers. Resources are specified via an rspec which is an XML formatted specification of the PCs and links required by the experiment. More information about Rspecs can be found at http://www.protogeni.net/trac/protogeni/wiki/RSpec
Additional information about using ProtoGENI can be found at http://www.protogeni.net/trac/protogeni/wiki/Tutorial
