Dear Department Heads:
As part of Yahoo!’s continuing effort to advance cloud and mobile computing research,
we are pleased to announce that, beginning in Summer 2010, we will make a cloud
computing cluster available to selected universities in the U.S. Below is a description of
our RFP process. The initial, two-page proposal is due on May 26, 2010. Please forward
this to your faculty members who might be interested.
Yahoo!’s M45 cluster is part of the Open Cirrus™ Testbed created by HP, Intel, and
Yahoo!, and the cluster has been in use by selected universities since November 2007.
The partition Yahoo! plans to make available to universities will consist of roughly 400
nodes with a total of 3,200 cores, over 2 terabytes of memory, and over 1 petabyte of raw
storage. The Yahoo! cluster is intended to run the Apache open source software Hadoop
and Pig. Each selected university will share the partition with other selected universities.
The initial duration of use is 6 months, potentially renewable for another 6 months upon
Our proposal evaluation process will consist of two rounds. In the first round, the
university will submit a short proposal. This proposal should be no more than 2 pages of
11-point text. The proposal shall consist of the following:
1. Name of the department and university (limit: one proposal per university)
2. Name, title/tenure status, research area, and brief bio of each PI (limit: 12 PIs per
3. Project description: For each PI, describe the proposed research project(s). The
proposal must describe why access to the computing power of the Yahoo! cluster is
necessary to advance this research. The projects should focus on exploring innovative
use of the cluster, running atop a Hadoop service, and research topics in cloud/data-
intensive computing. It is possible that systems data will be made available for analysis
and evaluation, and data can be shared across institutions. Note, however, that changes
to the operating environment will be limited, and request for access to lower layers of the
software stack and the hardware may be denied.
4. Personnel commitment: We request that each university provide one or more full-time
employees to provide dedicated technical/application-level support to the users at that
Please send the proposal to firstname.lastname@example.org with the subject line: PROPOSAL
from <name of university>. Deadline for proposal submission is 5 p.m., proposer’s local
time, on May 26, 2010.
We will evaluate proposals based on the following criteria:
1. Alignment of research interest. Of special interest is research that bridges mobile and
cloud computing, and novel use of cloud computing to support mobile services and
unique mobile user experiences. More broadly, we are interested in research projects in
areas that are aligned with Yahoo!’s business, including, but not limited to, the following:
extension/evaluation/experimentation of the Hadoop distributed file system, the Pig
parallel programming system, and related open source software; novel use of Hadoop
and Pig in new application domains; machine learning, search, computational advertising,
microeconomics and social systems, and community systems. Additional descriptions of
some of these areas of interest can be found at http://labs.yahoo.com/.
2. Strength of institution and track record of PIs in research areas of interest to Yahoo!.
3. Innovative use of the Yahoo! cluster to enable new kinds of research.
Universities with proposals moving to the second round will be notified by June 28, 2010
and will be invited to submit a more detailed proposal if need be. Due to the anticipated
volume of RFP responses, Yahoo! may not be able to notify universities that are not
invited to join the second round, although every effort will be made to do so. After the
second round, awardee universities will be required to negotiate and execute a usage
agreement with Yahoo! before access to the cluster will be granted. The agreement is
intended to cover all the PIs from that university. No access to the cluster will be granted
unless a properly executed usage agreement is in place.
Frequently Asked Questions
Q: How can we learn more about the cluster?
A: The cluster is described in the Yahoo!’s facility section in this RFP. Upon successful
execution of a usage agreement, additional information will be provided about the cluster.
Q: Is there any training offered on the cluster?
A: Yahoo! will provide training to the institution upon successful execution of a usage
Q: If I cannot reply by the RFP due date, will the RFP be re-opened in the future?
A: Yahoo! may reopen the RFP in the future but also reserves the right to not re-open.
Q: What does “share the partition” mean?
A: Users will log in and run code on the same, timeshared partition.
Q: Will others be able to see my application/data?
A: It is possible that researchers sharing the cluster may see each others applications
Q: Can one user’s application bring down the entire cluster?
A: The cluster is timeshared so this is possible but unlikely.
Q: What if I haven’t finished my research at the end of the 6 month period?
A: The initial duration of use is 6 months, potentially renewable for another 6 months
upon written agreement.
Q: How is Intellectual Property rights handled?
A: The policy regarding IP rights will be disclosed during the negotiation of the usage
Q: Where will we be able to find open sources tools for the cluster?
A: The cluster runs and supports Apache open-source software such as Hadoop and Pig.
Q: If we create new tools, will they all have to be open-sourced?
A: While we encourage researchers to contribute to the open-source community, we do
not require all new tools created to be open-sourced.