ISCA 2001 Tutorials (June 30 – July 1)

T1: VIA and InfiniBand Communication Architecture (half day)

Presenter: Dhabaleswar K. Panda (Ohio State University)

Saturday 9:00 AM – 12:30

This tutorial is intended to provide an in-depth overview of both Virtual Interface Architecture (VIA) and InfiniBand Architecture (IBA) standards. Issues related to designing high performance communication and I/O architectures for next generation scalable systems will be emphasized. Basic components of these standards will be discussed. New functionalities associated with the IBA standard such as Quality of Service, Protection Domains, etc., for designing large-scale enterprise servers will be discussed. Open research

challenges in designing scalable VIA and IBA implementations and providing support for higher-level programming models will be outlined. The tutorial will conclude with an overview of on-going VIA and InfiniBand related research at universities, industry, and research labs. For more information, see http://www.cis.ohio-state.edu/~panda/isca01_tut.html

T2: Virtual Machines: Architectures, Implementations, and Applications (half day)

Presenters: James E. Smith (University of Wisconsin) and Sriram Vajapeyam (IISc, Bangalore)

Saturday 1:30 PM – 5:00 PM

Virtual machines (VMs) -- program execution environments implemented via a combination of hardware and hidden software -- have emerged as a powerful means for tackling a number of important engineering problems, including software mobility and security, processor performance, and power optimization. We present an introduction to several hardware- software VM techniques, focusing primarily on combined hardware-software interaction. Topics to be covered include "classic" VMs, application VMs, platform-independent VMs (e.g. JavaVM), whole-system VMs, binary translation, and dynamic hardware optimization. Several case studies from both commercial implementations and research projects will be discussed. We conclude by discussing future directions for VM research. For more information, see http://www.csa.iisc.ernet.in/~sriram/isca01-tutorial/.

T3: Power-Efficient Design: Modeling and Optimizations (1 day)

Presenters: Pradip Bose (IBM), David Brooks (Princeton), Mary Jane Irwin (Penn State Univ.), Mahmut Kandemir (Penn State Univ), Margaret Martonosi (Princeton), and Narayanan Vijaykrishnan (Penn State Univ.)

Sunday 9:00 AM – 5:00 PM

With the rapid proliferation of battery-operated mobile systems and power-hungry high-performance systems, design for power efficiency has become of utmost importance. This tutorial will introduce the sources for the power consumption and identify the motivation for limiting the power consumption. Next, the tutorial will introduce the appropriate power and energy metrics to utilize based on the target system. Techniques for modeling energy consumption at different levels of accuracy will be illustrated. Following that both static and dynamic optimization techniques at the hardware and software levels will be presented. The talk will conclude with a discussion of future trends. For more information, see http://www.cse.psu.edu/~mdl/ or http://www.ee.princeton.edu/~mrm/tutorial

T4: Fault-tolerant CORBA (half day)

Presenter: Priya Narasimhan (Eternal Systems, Inc).

Sunday 9:00 AM – 12:30

In March 2000, the Object Management Group (OMG), the CORBA standards body, added reliability to the list of CORBA's attractive features by approving a new Fault-Tolerant CORBA standard. The tutorial will cover (i) the interfaces of the new standard, (ii) the requirements that a conformant fault-tolerant CORBA implementation must satisfy, (iii) the limitations of the standard, and (iv) some approaches for going beyond the standard, for instance, to cope with multithreaded CORBA applications. For more information, see http://www.dsn.org.

T5: System Survivability Analysis and Simulation (half day)

Presenters: Richard C. Linger, Nancy R. Mead, and David Fisher (Software Engineering Institute, Carnegie Mellon University)

Sunday 1:30 PM – 5:00 PM

Increasing dependence on critical systems amplifies the consequences of intrusions and compromises. Survivability is the capability of a system to fulfill its mission despite attacks, failures, and accidents. This tutorial describes the Survivable Network Analysis (SNA) method for assessing and improving survivability characteristics of systems. It also introduces Easel, a discrete event simulation system for simulating system survivability in general and emergent algorithms in particular. For more information, see http://www.dsn.org .