Although clusters are a popular form of high-performance computing
(HPC), they remain more difficult to manage than sequential systems,
or even symmetric multiprocessors.  Furthermore, as cluster sizes
increase, resource management---essentially, everything that runs on a
cluster other than the applications---becomes an increasingly large
impediment to application efficiency.  In this talk we present
STORM, a resource-management framework designed for scalability and
performance.  The key innovation behind STORM is a software
architecture that enables resource management to exploit low-level
network features.  As a result of this HPC-application-like design,
STORM is orders of magnitude faster than the best reported results in
the literature on two sample resource-management functions: job
launching and process scheduling.
Further, we identify a small set of network primitives that is
sufficient for a scalable implementation of a resource manager
if implemented itself in a scalable manner.