| Jochen Liedtke. Towards real microkernels. Communications of the ACM, 39(9):70--77, September 1996. |
.... recoverable and transactional memory [10, 45] provide more predictable or improved cache behavior via page coloring [28] enable predictable access timing via pinning, and even enable secure application controlled virtual memory by safely exporting control of basic virtual memory mechanisms [15, 35, 42]. We show that this powerful mechanism can also be used for kernel memory, and we demonstrate that similar benefits can be obtained. Clearly, if the kernel is to be paged by user level applications, care must be taken not to compromise protection, e.g. by allowing applications to see or even ....
Jochen Liedtke. Toward real microkernels. Communications of the ACM, 39(9), Sep 1996.
....17 5 Benchmarks 17 6 Future Work and Discussion 18 3 4 1 Introduction This document describes the mechanisms used within the L4 Alpha microkernel to extend the L4 API to systems with multiple processors. Basic knowledge of the L4 microkernel is assumed; for a more general discussion of L4, see [Lie95,Lie96], as well as the user manual [AH98] Basic knowledge of the Alpha family of microprocessors is also assumed; see [Com98] for more information. For an introduction to L4 Alpha see the L4 Alpha reference manual [PWH01] 1.1 Goals The development of this microkernel was done with the following ....
Jochen Liedtke. Towards real microkernels. Communications of the ACM, 39(9):70--77, September 1996.
.... executives to general purpose time sharing systems to be implemented on top of it, with the option for concurrently running other operating system personalities [HHW98] The biggest concern in using a microkernel base was the poor performance displayed by rst generation microkernels [CB93, Lie96d] The L4 microkernel [Lie95b, Lie96a, AH98] presented itself as an ideal choice not only due to its high performance implementations, but since the research group behind the PLEB project have a solid experience base with L4. This experience is due to the usage of L4 MIPS [EHL97] in teaching and ....
Jochen Liedtke. Towards real microkernels. Communications of the ACM, 39(9):70-77, September 1996.
....running. If the failed component was, say, the user interface on a machine running a web server, RR would allow availability of the web service to be unaffected. The ability to treat operating system services as separate components can avoid these failures, as evidenced by true microkernels [1, 24]. 3.2 Restartability Winners The classic replicated Internet server configuration has n instances of a server for a population of u users, with each server being able to handle in excess of u=n users. In such systems, node reboots result simply in a transient 1=n throughput loss. Moreover, ....
J. Liedtke. Toward real microkernels. Communications of the ACM, 39(9):70--77, 1996.
....malfunction in a service is isolated like a malfunction to a user program. Finally, a micro kernel system is more exible and tailor able: di erent implementations of the same service can easily co exist. A micro kernel design is far superior to a monolithic kernel in terms of software engineering [12, 13]. However, most micro kernels su er from poor performance, which has prevented their wide spread adoption. The exported procedure call (EPC) provides many of the above advantages without forcing the adoption of an entirely new operating system. There is some overhead to exporting an operation. ....
Jochen Liedtke. Toward real microkernels. Communications of the ACM, 39(9):70-77, September 1996.
....of the most frequently used operations. However, performance of these first generation microkernels proved disappointing, with applications generally experiencing a significant slowdown compared to a traditional ( monolithic ) operating system [CB93] Liedtke, however, has shown [Lie93, Lie95, Lie96] that these performance problems are not inherent in the microkernel concept and can be overcome by good design and implementation. L4 is the constructive proof of this theorem, as has been clearly demonstrated by Hartig et al. HHL 97] 1.1 L4 design philosophy The most fundamental task of ....
Jochen Liedtke. Towards real microkernels. Communications of the ACM, 39(9):70--77, September 1996.
....by modifying the DEC Ultrix 4.2A kernel, their approach requires only 8 microseconds for handling a null exception compared to 80 microseconds taken by the unmodified kernel. Another great example of fast exception handling mechanism is the L3 (and subsequently, L4) microkernel [Lie95, Lie96] the full cost of a kernel call in the L3 microkernel is between 123 and 180 cycles [Lie93] or less than one microsecond on a modern 200MHz processor. This is an extremely impressive result, given that our best performance on Linux is more than an order of magnitude slower. Anderson et al. ....
Jochen Liedtke. Toward Real Microkernels. Communications of the ACM, 39(9):70--77, September 1996.
....other. Minimal kernels Minimal operating systems can be easier to tune for performance, validate for reliability, and customize for multiple uses. Several groups have demonstrated significant performance improvements by implementing operating systems which are designed to be as small as possible [17, 18]. Other groups have built minimal kernels in order to demonstrate that it is possible to improve reliability and maintainability by proving (or at least arguing) that a (very) small core has been implemented correctly (see for example [26] and [10] A third set of groups has concentrated on ....
J. Liedtke. Toward real microkernels. Commun. ACM, 39(9):70--77, Sept. 1996.
....HADES base real time kernel so long as they offer enough functionalities to implement the HADES kernel adaption layer. The structuring principles that where adopted in HADES to meet the flexibility property are very similar to the ones proposed by micro kernel designers to specialize their kernels [16]. The difference comes mainly from the class of services that can be specialized. In micro kernels, services that can be specialized are general purpose operating system services (e.g. memory management specialization through paging services, communication services) while in HADES, such services ....
J. Liedtke. Toward real microkernels. Comm. of the ACM, 39(9):70--77, Sept. 1996.
....efficient low level communication layer of GAMMA may constitute a good support to higher level mechanisms of both message passing and shared memory parallel programming paradigms. Both could be implemented as user level libraries. Indeed this approach conforms to the current trend in OS technology [3, 5]: modern OS kernels have reduced sizes and functionalities, and many traditional OS services are moved to user level libraries, thanks to the modern technology of dynamic linking to shared libraries. In what follows, the words workstation and PC are used as synonyms. 2 Hardware configuration ....
J. Liedtke. Towards Real Microkernels. Comm. of ACM, 39(9):70--77, September 1996.
....has altered the scheduling status of a process. The selected process is then dispatched. Although we adopted a well known kernel architecture, a full kernel was not implemented. We refer to the implemented subset as a nanokernel. Although the term has been disparaged, we believe it is warranted [Lie96]. Although similar to a microkernel, a nanokernel has the following identifying characteristics: ffl There exists only one mandatory nanokernel API routine: run a specified subroutine serialized in kernel context, that is, as a fork routine. There are no other API s defined by the nanokernel. ....
Jochen Liedtke. Towards real microkernels. Communications of the ACM, 39(9):70--77, September 1996.
No context found.
Jochen Liedtke. Towards real microkernels. Communications of the ACM, 39(9):70--77, September 1996.
No context found.
J. Liedtke. Towards real microkernels. Communications of the ACM, 39(9):70--77, Sept. 1996.
No context found.
J. Liedtke. Towards real microkernels. CACM, 39(9):70--77, Sep 1996.
No context found.
J. Liedtke. Toward real microkernels. Communications of the ACM, 39(9):70, September 1996.
No context found.
Jochen Liedtke. Towards real microkernels. Communications of the ACM, 39(9):70--77, September 1996.
No context found.
Liedtke, J.: Toward real microkernels. Communications of the ACM 39 (1996)
No context found.
Jochen Liedtke. Toward real microkernels. Communications of the ACM, 39(9), Sep 1996.
No context found.
J. Liedtke. Toward real microkernels. Communications of the ACM, 39(9):70--77, 1996.
No context found.
J. Liedtke. Toward real microkernels. Communications of the ACM, 39(9):70--77, 1996.
No context found.
Jochen Liedtke. Toward real microkernels. Communications of the ACM, 39(9):70--77, 1996.
No context found.
J. Liedtke. Toward Real Microkernels. Comm. of the ACM, 39(9):70--77, Sept. 1996.
No context found.
Jochen Liedtke. Toward real microkernels. Communications of the ACM, 39(9), Sep 1996.
No context found.
Jochen Liedtke. Toward real microkernels. Communications of the ACM, 39(9), Sep 1996.
No context found.
J. Liedtke, "Toward Real Microkernels," Communications of the ACM, vol. 39, no. 9, Sep. 1996.
No context found.
Liedtke, J., "Toward Real Microkernels". Communications of the ACM (CACM), 39(9), pp 70-77, September 1996.
No context found.
Liedtke, J.: Toward Real Microkernels, Communication of the ACM, Vol. 39, No. 9, September 1996, pp. 7077
No context found.
Liedtke, J., "Toward Real Microkernels". Communications of the ACM (CACM), 39(9), pp 70-77, September 1996.
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