Jon Inouye

Optimistic incremental specialization: streamlining a commercial operating system

Conventional operating system code is written to deal with all possible system states, and performs considerable interpretati on to determine the current system state before taking action. A consequence of this approach is that kernel calls which perform little ac tual work take a long time to execute. To address this problem, we use specialized operating system code that reduces interpretation for common cases, but still behaves correctly in the fully general c ase. We describe how specialized operating system code can be generated and bound incrementallyas the information on which it depends becomes available. We extend our specialization techniques to include the notion of optimistic incremental specialization : a technique for generating specialized kernel code optimistically for sys tem states that are likely to occur, but not certain. The ideas outlined in this paper allow the conventional kernel design tenet of “optimi zing for the common case” to be extended to the domain of adaptive operating systems. We also show that aggressive use of specialization can produce in-kernel implementations of operating system functionality with performance comparable to user-level implementations. We demonstrate that these ideas are applicable in real-world operating systems by describing a re-implementation of the HP-UX file system. Our specializedread system call reduces the cost of a single byte read by a factor of 3, and an 8 KB read by 26%, while preserving the semantics of the HP-UXread call. By relaxing the semantics of HP-UXread we were able to cut the cost of a single byte read system call by more than an order of magnitude.

Dynamic network reconfiguration support for mobile computers

Hot swapping technology combined with pervasive heterogeneous networks empowers mobile laptop users to select the best network device for their current environment. Unfortunately, the majority of system software remains “customized”for a particular network configuration, and assumes that many characteristics associated with the network environment remain invariant over the runtime of the software. Mobility causes changes in the environment and nullifies many of these assumptions. This leads to severe loss in system functionality when resources are lost, and failure to benefit when resources are gained. This paper describes Physical Media Independence (PMI), an architecture for dynamically diverse network interface management. PMI addresses three issues concerning dynamic network configuration. First, a model for device availability is proposed to accurately determine when a network device is operational. Second, a structured methodology is used to construct adapters that reconfigure the system when the set of available devices change. The methodology breaks traditional layering using a meta-interface to pass events and information among layers, allowing each layer to adapt in a manner best suited to it. Third, we examine the effect of transparent and non-transparent reconfiguration operations on a variety of applications. We find that adaptive, resource intensive applications behave more efficiently when informed of device availability events. We compare the functionality of an adaptive application running on top of a adaptive operating system (OS) with and without cross-layer notifications.

The effects of virtually addressed caches on virtual memory design and performance

Recent times have witnessed rapid advances in microprocessor technology resulting in an order of magnitude performance improvement every few years. These developments in hardware have been paralleled by several prominent trends in operating system design, the most notable being a move towards message-passing micro-kernels. However, operating system performance has not kept pace with that of the underlying hardware. It has become apparent that design changes to enhance processor performance can have adverse effects on operating system performance. This problem arises when the architectural assumptions implicit in an operating systems design are inappropriate for the architectures on which it executes.This paper examines one specific area in which operating system design assumptions appear to be in conflict with trends in modern processor architecture. We focus on the performance effects of virtually addressed caches on two contemporary operating systems (Mach and Chorus). We present experimental results to illustrate the impact of virtually addressed caches on the performance of primitive virtual memory operations, and higher-level operations, such as inter-process communication, that utilize these primitive operations. The main goal of the paper is to encourage operating system designers to revisit some of the basic architectural assumptions implicit in modern operating system designs.