The Anonymous Remote Computing Project
The project on Anonymous Remote Computing paradigm is about an
environment for parallel and
distributed computing on workstation clusters in presence of heterogeneity,
failures and varying load.
The focus is on separation of concerns, i.e. separation of system's
concerns from programmer's concerns.
A parallel program is written in an extended C supporting ARC function
calls, which can have blocking or non-blocking
semantics. An ARC function call is executed by the ARC kernel
on a remote machine without the involvement of the
programmer. Processors can join or leave the system at any time and
are employed by the kernel as and when
required. The user can obtain a number called horse power factor
(HPF) that indicates the current capacity
of an anonymous machine. The HPF normalizes the load and speed
of machines. Source code of an ARC function
is migrated and compiled in case of heterogeneous architectures.
Originally an ARC framework was written on a heterogeneous cluster consisting
of RS6000s and Sun3s and Sun4s,
and then on a Linux cluster.
Currently it's also available for C# over .NET platform.
Link to Implementation Pages
List of Students Who have contributed to this project, or are currently
working on ARC:
Yamini Sharma, Kalpesh Kapoor, Satish Umale, Mauresh Phadke,
L. Aruna,
Vamsi Kalyan, Amit Soni, Satyashil Avadhare
Documentation on ARC
Code and Download an ARC Implementation
ARC Related ideas can be read through following papers
Rushikesh K.Joshi, D. Janaki Ram,
Anonymous Remote Computing: A Paradigm for
Parallel Programming on Interconnected Workstations,
IEEE Transactions on Software Engineering, Jan/Feb 1999
Aruna L., Yamini Sharma, Rushikesh K. Joshi,
Design and Implementation of an RPC based ARC Kernel,
Proceedings of HPCN 2001,
LNCS Vol. 2110, pp.251-262.
R.K. Joshi, D Janaki Ram,
Object Based Subcontracting for Parallel
Programming on
Loosely-Coupled Distributed Systems,
Journal of Programming Languages, 4 (1996), pp. 169-183.
Rushikesh K. Joshi, D. Janaki Ram,
Parset: A language construct for system
independent
parallel programming on distributed systems,
Microprocessing and Microprogramming (now the Journal of
Systems Architecture), June 1995, pp. 245-259.
T. Vamsi Kalyan, Rushikesh K. Joshi,
Architecture of the Object Oriented Anonymous Remote Computing
Framework for C# over .NET
Presented at 2nd Workshop on Software Design and Architecture (SoDA),
Jan 2001, Bangalore.