J. Hennessy. The Future of Systems Research. Computer, 32(8):27--33, August 1999.  Home/Search   Document Not in Database   Summary   Related Articles   Check

.... has gained momentum in recent years with the ever increasing complexity of these systems in fact, several researchers have argued that, in today s environments, the problems of manageability, availability and incremental growth have overshadowed that of the traditional emphasis on performance [12, 18]. 1.2 Research Contributions This paper focuses on the architecture of a selfmanaging web server that supports multiple QoS classes a scenario where multiple virtual servers run on a single physical server or where certain classes of customers are given preferential service. Assuming such an ....

J. Hennessy. The Future of Systems Research. IEEE Computer, pages 27--33, August 1999.

....over 60 embedded processors that control a multitude of functions, e.g. the fuel injection and the anti lock braking system (ABS) that guarantee a smooth and foremost safe drive. The employment of embedded processors appear to grow in an exponential curve. Furthermore, it has been postulated [22] that the sales trend of embedded processors (microprocessors in this setting) will significantly outperform the sales of general purpose PC processors. In this positional paper, we describe several characteristics of embedded processors and investigate how these characteristics have changed over ....

J. Hennessy. The Future of Systems Research. Computer, pages 27--33, 1999.

....argued the same point, insisting that performance should be less of an emphasis. Instead, other qualities will become crucial: availability [ maintainability, and] scalability. For servers if access to services on servers is the killer app availabil ity is the key metric [23]. Furthermore, the traditional scalability problem, of creating and efficiently using large massively parallel systems, is giving way to what we call the evolutionary growth problem : constructing large scale servers that can be incrementally expanded using newer, heterogeneous components. ....

....will naturally follow. As part of the Berkeley ISTORE project [9] we have taken on the challenge of effecting this transition by building reproducible, cross platform AME benchmarks for Availability, Maintainability, and Evolutionary Growth, the three challenge areas laid out by Hennessy [23]. This report presents our first steps toward that goal. We have chosen to focus initially on availability and maintainability, as we suspect that the task of measuring evolutionary growth boils down to measuring availability and maintainability as system scale is increased. We have developed ....

J. Hennessy. The Future of Systems Research. IEEE Computer 32(8):27--33, August 1999.

....messaging between mobile hosts without a central server. The main motivation to develop a new system stemmed from a fundamental and quantifiable shift in the global computing landscape from desktop and centralized mainframe operating systems towards mobile, embedded and wireless systems [5]. Dubbed ubiquitous computing [13] or the Post PC environment, this trend towards embedding computers in everyday devices shifts the focus of system development away from multi user timesharing issues towards distributed systems. While the impact of these changes on the theoretical course ....

J. Hennessy. The Future of Systems Research. In IEEE Computer, pages 27-33, August 1999.

....Trouble Free Systems, which can largely manage themselves while providing a service for millions of people. Butler Lampson [1999] has called for systems that work: they meet their specs, are always available, adapt to changing environment, evolve while they run, and grow without practical limit. Hennessy [1999] has proposed a new research target: availability, maintainability, and scalability. IBM Research [2001] has announced a new program in Autonomic Computing, whereby they try to make systems smarter about managing themselves rather than just faster. Finally, Bill Gates [2002] has set trustworthy ....

J. Hennessy, The Future of Systems Research, Computer, August 1999 32:8, 27-33.

....been demonstrated to achieve significant performance benefits for traditional engineering and scientific workloads. Consequently, the use of large caches is a common trend across general purpose systems, sometimes consuming up to 80 of the total transistor budget and up to 50 of the die area [46]. Below, we first discuss the impact of varying the size of the second level cache (Section 2.7.1) and then discuss the impact of varying the size of the first level cache (Section 2.7.2) Section 2.7.3 summarizes our results. 32 2.7.1 Impact of Varying Second Level Cache Sizes Overall results. ....

....this philosophy to the design of caches. We focus on caches because the use of large caches is a common trend across current general purpose systems, typically consuming the largest fraction on the on chip transistor resources (up to 80 of the total transistor budget and up to 50 of the die area [46]) Additionally, as discussed in Chapter 2, while large caches are effective for a variety of conventional workloads, they are often ineffective for media processing applications because of the streaming nature of data accesses, the large working sets, and the compute bound nature of these ....

John Hennessy. The Future of Systems Research. IEEE Computer, 32(8):27--33, August 1999.

....the need for design trade offs. In the August issue of IEEE Computer the provost of Stanford John Hennessy writes Performance long the centerpiece needs to share the spotlight with availability, maintainability, and other quality attributes in his article on The Future of System Research [11]. The Architecture Trade off Analysis Method (ATAM) proposed by Kazman et al. 14] addresses the need for making trade offs between different quality attributes. One important difference between our guidelines and ATAM is that the ATAM work concentrates on identifying so called trade off points, ....

J. Hennessy, "The Future of System Research", IEEE Computer, August 1999.

....The point at which increasing bandwidth does not improve performance occurs at a bandwidth much greater (by a factor of 5) for snooping than for directories. Designing a single protocol to provide high performance for many system configurations and workloads is difficult. Hennessy writes [14], W]e don t have a coherency scheme that does well under all these situations: from small to large processor counts, different levels of [software] optimization, and differing cache sizes. We advocate an adaptive approach to address this challenge. There are two reasons why an adaptive scheme ....

J. Hennessy. The Future of Systems Research. IEEE Computer, 32(8):27--33, Aug. 1999.

....a gradual growth. Some of the driving forces behind the fast expansion of the embedded systems market relates to 700 600 500 400 300 200 100 0 1996 1997 1998 1999 2000 2001 2002 2003 Units shipped (millions) Embedded processors PC processors Figure 1: Marketing projections. from [1]) the proliferation of computing technologies to traditionally non computing domains, e.g. medical instrumentation, automotive industry, the fast advances in VLSI technologies, and the tendency to replace analog signal processing with digital signal processing. In this paper, we discuss ....

J. Hennessy, "The Future of Systems Research," Computer, pp. 27--33, August 1999.

....more CPU power, one can buy processing nodes. 1.1 The Problem: Management However, storage clusters also introduce additional challenges, particularly regarding manageability. Whereas absolute performance was the goal of a great number of previous systems, manageability has become the new focus [16, 25]. Thus, a system that works consistently with little or no human intervention will be preferred over a system that sporadically delivers near peak performance or requires a large amount of human attention to do so. Manageability is more challenging in storage clusters due to their additional ....

J. L. Hennessy. The Future of Systems Research. IEEE Computer, 32(8):27--33, August 1999.

....quality service in the presence of errors that can affect the system in operation. A recent article in IEEE Computer observes that the computer industry and computer research must focus on availability, maintainability and scalability of computer systems; performance should be less of an emphasis [Hennessy 99] In addition, for emerging future technologies like molecular computing [Collier 99] dependability will be a major cause of concern. This concern has been expressed in many articles on molecular computing, including [Peterson 00] Quinlan 99] Wall Street 99] Some of the major components of ....

Hennessy, J., "The Future of Systems Research," IEEE Computer, Vol. 32, No. 8, pp. 27-33, August 1999.

....for the design of interface for Internet TV. Keywords: Internet TV, Internet Appliance, Convergence, Converging anchor, Converged value, Theory of convergence 1 Introduction Recently, computing environment is changing from the era of personal computer (PC) to the age of post desktop (Hennessy, 1999). Post desktop means that the placement of computer is no longer limited to a desk. Network computers will be accessible to everyone and everywhere in this new age. For example, a housekeeper can get a recipe in the kitchen through an Internet refrigerator or a family can buy a piece of furniture ....

Hennessy, J. (1999), The Future of Systems Research, Computer, 32(8), 27-33.

....superscalars use the rising transistor counts in extracting instruction level parallelism (ILP) to achieve high performance. Unfortunately, superscalar architectures are not only becoming less effective in improving the clock speed [25, 21,1] and ILP but also worsening in design complexity [20] and reliability [2] across chip generations. Instead, many researchers and vendors are exploiting the increasing number of transistors to build chip multiprocessors (CMPs) by partitioning a chip into multiple simple ILP cores [29,18] As in traditional multiprocessors, CMPs extract thread level ....

J. Hennessy. The future of systems research. IEEE Computer, 32(8):27--33, Aug. 1999.

....parallelism and speculative out of order execution. While this trend has led to significant performance gains on target applications such as the SPEC benchmark [40] continuing along this path is becoming less viable due to substantial increases in development team sizes and design times [18]. Furthermore, more complex designs are yielding diminishing returns in performance even for applications such as SPEC. Meanwhile, commercial workloads such as databases and Web applications have surpassed technical workloads to become the largest and fastest growing market segment for ....

J. Hennessy. The Future of Systems Research. In IEEE Computer, Vol. 32, No. 8, pages 27-33, August 1999.

....our computing infrastructure. Kimera A Scalable and Secure System Architecture for Networked Computers In my thesis work[13] I address the security, manageability and scalability problems faced by networked computers. The rise of the Internet, the explosion in the number of networked hosts [6], and the accompanying shift from desktop computing on workstations to network computing on ubiquitous devices [17] have outpaced the capabilities of modern operating systems. In particular, even though the Java virtual machine was speci cally designed to o er strong security guarantees, to ....

John Hennessy. The Future of Systems Research. IEEE Computer, pages 27-33, August 1999.

....are implemented in a rigid manner and their resources are wasted for applications that cannot directly utilize them. For example, the use of large caches is a common trend across general purpose systems, sometimes consuming up to 80 of the total transistor budget and up to 50 of the die area [9]. While large caches are effective for a variety of conventional workloads, they are often ineffective for media processing applications because of the streaming nature of data accesses and the large working sets in these applications [19] An alternative design philosophy is to build some ....

J. Hennessy. The Future of Systems Research. IEEE Computer, 32(8):27--33, August 1999.

....the area of mobile computing. 4 Systems research Within the research community there is a growing recognition of the fact that systems research can no longer focus almost exclusively on performance but must shift its attention towards the end user requirements for dependability and ease of use [16]. This need is even more acute in the mobility field where the most visible impact of software engineering research will be in the wide range of applications expected to emerge on the market in the very near future. This suggests a need to consider a style of research that is much more application ....

J. Hennessy. The Future of Systems Research. Computer, 32(8):27--33, August 1999.

No context found.

J. Hennessy. The Future of Systems Research. Computer, 32(8):27--33, August 1999.

No context found.

J. Hennessy, "The Future of Systems Research," Computer, Aug. 1999, pp. 27-33.

No context found.

J. Hennessy. The Future of Systems Research. IEEE Computer 32(8):27--33, August 1999.

No context found.

J. Hennessy, "The Future of Systems Research," Computer, Aug. 1999, pp. 27-33.

No context found.

J. Hennessy. The Future of Systems Research. IEEE Computer, pages 27--33, August 1999.

No context found.

J. Hennessy. The Future of Systems Research. In IEEE Computer, pages 27--33, August 1999.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC