Marc E. Fiuczynski

Container-based operating system virtualization: a scalable, high-performance alternative to hypervisors

Hypervisors, popularized by Xen and VMware, are quickly becoming commodity. They are appropriate for many usage scenarios, but there are scenarios that require system virtualization with high degrees of both isolation and efficiency. Examples include HPC clusters, the Grid, hosting centers, and PlanetLab. We present an alternative to hypervisors that is better suited to such scenarios. The approach is a synthesis of prior work on resource containers and security containers applied to general-purpose, time-shared operating systems. Examples of such container-based systems include Solaris 10, Virtuozzo for Linux, and Linux-VServer. As a representative instance of container-based systems, this paper describes the design and implementation of Linux-VServer. In addition, it contrasts the architecture of Linux-VServer with current generations of Xen, and shows how Linux-VServer provides comparable support for isolation and superior system efficiency.

On the feasibility of an AOSD approach to Linux kernel extensions

In previous work, we presented a domain-specific version of C, called C4, which was used for capturing extensions to the Linux 2.6 kernel using AOSD techniques as an alternative to the conventional patching approach [10, 19]. The focus of that work was on introducing new extensions represented as aspects in system software such as the Linux kernel with a focus on readablility, compatibility, performance, and the preservation of existing development workflows. However, other AOSD researchers (e.g. Lohmann et al. [8]) state that ... Linux, as a monolithic system, provides a low number of join-points for aspects and that those available were semantically ambiguous. This worrisome statement motivated us to study the feasibility of applying AOSD techniques to refactor existing Linux kernel extensions. To gain insight we analyzed the AOSD-ness of a large number of configurable options available in the Linux kernel and evaluated whether they could be converted into aspects---and for the AOSD fan the preliminary results are promising.

PlanetLab: overview, history, and future directions

The first OSR issue I read was Volume 29 No. 2 (April 1995). At the time I was actively working on the SPIN Extensible Operating System [5] and looking for new ways to safely expose interrupt handling to user-provided code loaded into the kernel. The article by Kleiman and Eykholt on Interrupts as Threads in that issue of OSR was highly relevant and informative. Ever since then I had planned on contributing to OSR but unfortunately never got around to it. For this reason, when Jeanna Matthews asked whether I could serve as the guest editor of the inaugural special topics OSR issue, I jumped at the opportunity.

Privileged operations in the PlanetLab virtualised environment

Virtualised systems have experienced a resurgence in popularity in recent years, whether used to support multiple OSes running on a users desktop, provide commercial application hosting facilities, or isolate a large number of users from each other in global network testbeds. We also see an increasing level of interest in having entities within these virtualised systems interact with each other, either as peers or as helpers providing a service to clients.Very little work has been previously conducted on how such interaction between virtualised environments can take place. We introduce Proper, a service running on the PlanetLab system, that allows unprivileged entities to access privileged operations in a safe, tightly controlled manner.This paper describes our work designing and implementing Proper, including a discussion of the various architectural decisions made. We describe how implementing such a system in a traditional UNIX environment is non-trivial, and provide a number of examples of how services running on PlanetLab actually use Proper.

IPv6 Translation for IPv4 Embedded Systems

The US Department of Defense (DoD) plans to add Internet protocol (IP) version 6 (IPv6) capability to all DoD IP networks by 2008 then begin a phase-out of IP version 4 (IPv4). To facilitate this migration various transition mechanisms have been developed to address interoperability of IPv4 and IPv6 networks and systems. Unfortunately, none of the existing mechanisms address two fundamental problems: (1) noninvasive migration of essential legacy IPv4-only systems to IPv6, and (2) operation of IPv4-only systems on IPv6-only core networks. This paper presents a solution to these problems by using a simple, low cost IPv4/IPv6 proxy translation device

On reconciling patches and aspects

In previous work, we presented a domain-specific enhancement to C, called C4, that lets developers manage program extensions leveraging techniques inspired by the AOSD model as an alternative to the conventional patch approach [3]. Our goal is to offer: (1) tool compatibility letting programmers develop, integrate, modify, and debug C4-based extensions that preserve their existing development workflow and leverages their existing tools rather than requiring additional tools; (2) code understandability of the C4 syntax such that is it straightforward for an uninitiated C programmer to use immediately; and, (3) runtime performance achieving near-zero overhead such that it can be used even in performance critical execution paths. As such C4 source code can be viewed as the result of weaving in AOSD style introductions and advices inline into C program. However, C4 lacked a proper representation of its unwoven form--i.e., whats conventionally in AOSD circles referred to as the pointcut language. This paper makes a case for B4: a patch-based pointcut representation of unwoven C4 and contrasts it with development-oriented pointcut languages belonging to the AspectC family that have been defined for the C programming language.

Networking Technologies in the Developing World

In this paper, we describe the networking and distributed systems challenges and opportunities that are present in adoption of information and computing technology (ICT) in the developing world, and illustrate connections between these challenges and current distributed systems and networking research programs. We suggest that this agenda form a core element of the ambitious program of network experimentation now commencing in the United States, European Union, and Japan.

Minding the gap: R&D in the Linux kernel

The Linux kernel, since its inception in 1991, has captured the interest of many thousands of developers and millions of users. It recently celebrated its 16th anniversary, includes many millions of lines of code, and is used in production systems around the world. It is also advancing at an increasingly rapid pace, undergoing many changes every single day. Indeed the kernels importance to many large corporations has sparked a high level of contribution by those companies [3] [4], including the employment of many core kernel developers. Recently Linus Torvalds published statistics relating to contributions to the kernel over the past three years: 96,885 patches attributed to 4068 distinct authors have been accepted [5].