X. Liu, C. Kreitz, R. van Renesse, J. Hickey, M. Hayden, K. Birman, and R. Constable. Building Reliable, HighPerformance Communication Systems from Components. In Symposium on Operating Systems Principles, 1999.  Home/Search   Document Not in Database   Summary   ACM   TOC   Related Articles   Check

....This approach allows assume guarantee reasoning to prove properties of compositions. To fit with this approach, an Amalia generated layer would need to be specified as ASM fragments and required provided properties. There are interesting parallels between our approach and that of Liu et.al. [11] on correctness of component compositions and synthesis of optimized code from these compositions. They start with a set of micro protocols (specialized components) which are written in a formally defined dialect of ML. A tool compiles the microprotocols into functions that are parameterized by ....

Xiaoming Liu, Christoph Kreitz, Robbert van Renesse, Jason Hickey, Mark Hayden, Kenneth Birman, and Robert Constable. Bulding reliable, high-performance communication systems from components. Operating Systems Review, 34(5), 1999. SOSP-17.

....of the type of pairs of integer numbers (which would represent the numerator and the denominator of a fraction) with the appropriate equality predicate. Somewhat surprisingly, it turns out that formulating rules for these quotient types is far from being trivial and numerous applications of NuPRL [6,19] have run into difficulties. Often a definition involving a quotient will look plausible, but after some (sometimes substantial) work it turns out that a key property is unprovable, or false. This work was partially supported by AFRL grant F49620 00 1 0209 and ONR grant N00014 01 1 0765. V.A. ....

Xiaoming Liu, Christoph Kreitz, Robbert van Renesse, Jason J. Hickey, Mark Hayden, Kenneth Bitman, and Robert Constable. Building reliable, high-performance communication systems from components. In 17 tn A CM Symposium on Operating Systems Principles, December 1999.

....Other researchers have used specialization in conjunction with specially designed system components to optimize a specific operating system service. For example, the Ensemble system uses specialization of network stacks written in ML in order to achieve high performance from modular components [32]. The Scout operating system achieves high performance by flattening network stacks automatically based on a new abstraction, paths, 27 which are defined by programmers [38] Scout uses domain specific language compilers to produce optimized code from the path specifications. 7 ....

Liu, X., C. Kreitz, R.v. Renesse, J. Hickey, M. Hayden, K. Birman, and R. Constable. Building reliable, high-performance communication systems from components. in 17th ACM Symposium on Operating Systems Principles. 1999.

....[7] which provides a collection of (relatively coarse grained, COM like) components implementing typical OS functionalities. The OSKit components have been used to implement several highly specialized OSes, such as implementations of the programming languages SML and Java at the hardware level [15]. OSKit components can be statically configured using the Knit tool [24] The Knit compiler modifies the source code to replace calls across component boundaries by direct calls, thus enabling standard compiler optimizations. Unlike THINK however, OSKit does not provide a framework for binding ....

....kernel functionality, namely a resource management framework and a communication framework. The resource management framework is original, whereas the communication framework is inspired by the x kernel [11] Other operating system level component based frameworks include Click [18] Ensemble [15] and Scout [19] These frameworks, however, are more specialized than THINK or OSKit: Click targets the construction of modular routers, Ensemble and Scout target the construction of communication protocol stacks. We thus believe that THINK is unique in its introduction and systematic application ....

[Article contains additional citation context not shown here]

X. Liu, C. Kreitz, R. van Renesse, J. Hickey, M. Hayden, K. Birman, and R. Constable. Building reliable, highperformance communication systems from components. In SOSP'99 [30].

....can be used to check the specification in this case. Second, the fast path involves code that is extracted from different processes and thereby makes some scheduling decisions. The ESP compiler has to ensure that it does not introduce starvation in the program. A number of research projects [85, 75, 68] investigate the use of fast paths in sequential programs. The Synthetix project [85] manually generated fast paths in the HP UX operating system. They show that file system calls like read can be speeded up by generating specialized code for it based on invariants and quasi invariants that are ....

....optimize these paths. To address this, the compiler generates optimized code for some paths. When a path is created at runtime, if the runtime system can determine that optimized code is available for that path, it uses the optimized code. Formal methods can be used to build optimized fast paths [68] in the Ensemble network architecture [51] A protocol stack consists of a sequence of protocol layers. The NuPRL [34] system was used to semiautomatically extract a fast path from the protocol stack. It ensures that the fast path generated is provably semantically equal to the protocol stack when ....

X. Liu, C. Kreitz, R. van Renesse, J. Hickey, M. Hayden, K. Birman, and R. Constable. Building Reliable, High-Performance Communication Systems from Components. In Proceedings of the Symposium on Operating Systems Principles, Kiawah Island Resort, South Carolina, December 1999.

....Furthermore, the use of fine grained components with simple specifications can make a system easier to understand. Because of its requirements for flexibility and extensibility, and its di#culty, networking software has been a popular field for the application of component techniques [6, 8, 9, 11, 13, 15]. Networks get faster at an even greater rate than processors, however, making the e#ciency of networking software ever more important. Even as component systems make networking software easier to program, component techniques introduce ine#ciencies that monolithic softPermission to make digital ....

....be easily extracted from these systems, so optimizing a configuration must proceed piecemeal. We know of no automatic optimizers for these systems. Previousworkapplyingcompileroptimizations tocomponents and systems tends to be more integrated with traditional compiler technology. The Ensemble [9] and Esterel [4] projects, for example, use language processors thatunderstandtheentiretyof theprogramming languages used to write their systems components. In contrast, our optimizers understand components at a high level; some do not handle C , the implementation language, at all. The ....

[Article contains additional citation context not shown here]

Xiaoming Liu, Christoph Kreitz, Robbert van Renesse, Jason Hickey, Mark Hayden, Kenneth Birman, and Robert Constable. Building reliable, high-performance communication systems from components. In Proc. 17th ACM Symposium on Operating Systems Principles (SOSP), pages 80--92, December 1999.

....of the type of pairs of integer numbers (which would represent a numerator and a denominator of a fraction) with the appropriate equality predicate. Somewhat surprisingly, it turns out that formulating rules for these quotient types is far from being trivial and numerous applications of NuPRL [6,16] have run into diculties. Often a de nition involving a quotient will look plausible, This work was partially supported by AFRL grant F49620 00 1 0209 and ONR grant N00014 01 1 0765 Cornell University Technical Report TR2002 1869, April 2002 but after some (sometimes substantial) work it ....

Xiaoming Liu, Christoph Kreitz, Robbert van Renesse, Jason J. Hickey, Mark Hayden, Kenneth Birman, and Robert Constable. Building reliable, high-performance communication systems from components. In 17 th ACM Symposium on Operating Systems Principles, December 1999.

....With testing it has in common that it assists in finding programming errors. In the VFiasco project our concern is not so much to find errors, but to give guarantees about their absence. Theorem proving. There are a few projects that apply theorem proving at the source code level as we do. In [7] Liu and colleagues use the theorem prover Nuprl to verify the correctness of network protocol stacks and to optimize such stacks. There are two major differences to the VFiasco project. First, to enable the verification the original C source code was rewritten in a carefully chosen subset of the ....

X. Liu, C. Kreitz, R. van Renesse, J. Hickey, M. Hayden, K. P. Birman, and R. L. Constable. Building reliable, high-performance communication systems from components. In 17th ACM Symposium on Operating System Principles (SOSP), pages 80--92, Kiawah Island, SC, December 1999.

....source code verification in combination with a well tested compiler provides enough security for all practical purposes. Theorem proving. Related work that comes closest to our project is those that applies theorem proving to source code and more especially to operating systems source code. In [17] Liu and colleagues use the theorem prover Nuprl [3] in the Ensemble project to verify the correctness of network protocol stacks and to optimize such stacks. They use several tools that translate their specifications and the Ensemble source code into Nuprl and back. In the VFiasco project we also ....

X. Liu, C. Kreitz, R. van Renesse, J. Hickey, M. Hayden, K. P. Birman, and R. L. Constable. Building reliable, high-performance communication systems from components. In 17th ACM Symposium on Operating System Principles (SOSP), pages 80--92, Kiawah Island, SC, December 1999.

....of previously known approaches, while avoiding most of their disadvantages. 1 Introduction Over the last decade record types and especially the dependent record types (or more formally dependently typed records) proved to be an extremely useful program speci cations and veri cation tool [7, 3, 11]. There are several approaches of de ning record types. Probably the most straightforward way of adding records to the type theory is to add them as a new primitive type. This approach was taken for example in [8] Unfortunately, if this approach is taken, the theory needs to be extended with the ....

Xiaoming Liu, Christoph Kreitz, Robbert van Renesse, Jason J. Hickey, Mark Hayden, Kenneth Birman, and Robert Constable. Building reliable, high-performance communication systems from components. In 17 th ACM Symposium on Operating Systems Principles, December 1999.

No context found.

Xiaoming Liu, Christoph Kreitz, Robbert van Renesse, Jason Hickey, Mark Hayden, Kenneth Birman, and Robert Constable. Building reliable, high-performance communication systems from components. In 17 th ACM Symposium on Operating Systems Principles (SOSP'99), pages 80--92, 1999.

No context found.

Xiaoming Liu, Christoph Kreitz, Robbert van Renesse, Jason J. Hickey, Mark Hayden, Kenneth Birman, and Robert Constable. Building reliable, high-performance communication systems from components. In 17 ACM Symposium on Operating Systems Principles (SOSP'99), volume 33(5) of Operating Systems Review, pages 80--92. ACM Press, December 1999.

No context found.

Xiaoming Liu, Christoph Kreitz, Robbert van Renesse, Jason J. Hickey, Mark Hayden, Kenneth Birman, and Robert Constable. Building reliable, high-performance communication systems from components. In 17 ACM Symposium on Operating Systems Principles (SOSP'99), volume 34 of Operating Systems Review, pages 80-92, December 1999.

No context found.

X. Liu, C. Kreitz, R. van Renesse, J Hickey, MHayden, K. Birman, R. Constable. Building reliable, high-performance communication systems from components. SOSP'99, Operating Systems Review 34(5):80-92, 1999.

....Group Manager The Group Manager is the key part of the distribution side of the file system. It determines what computers are part of the distributed MojaveFS system. The Group Manager also provides a form of reliable, totally ordered broadcast among group members, similar to the Ensemble System [8, 5]. Each node in the file system keeps a list of nodes it can communicate with, called the view. The view is changed whenever a node joins or leaves the group; this is called installing the view. The properties of group communication are best described through the following invariants. Self If ....

Xiaoming Liu, Christoph Kreitz, Robbert van Renesse, Jason Hickey, Mark Hayden, Kenneth Birman, and Robert Constable. Building reliable, highperformance communication systems from components. In 17 ACM Symposium on Operating Systems Principles (SOSP'99), volume 34 of Operating Systems Review, pages 80--92, 1999.

....(such as parsing, type inference, and scoping) that is assigned to the compiler is not available to the formal system, hindering e ective formal software development. In another related e ort, we used the NuPRL system to optimize communication protocols for the Ensemble group communication system [12, 11]. Again, this project separated the prover from the compiler. To optimize a protocol, a parser would convert the protocol and requirements into an expression in the NuPRL type theory; the prover would apply optimization tactics to generate a fast path; and the result would be printed as a ML le ....

Xiaoming Liu, Christoph Kreitz, Robbert van Renesse, Jason Hickey, Mark Hayden, Kenneth Birman, and Robert Constable. Building reliable, high-performance communication systems from components. In 17 ACM Symposium on Operating Systems Principles (SOSP'99), volume 34 of Operating Systems Review, pages 80-92, 1999.

No context found.

Xiaoming Liu, Christoph Kreitz, Robbert van Renesse, Jason Hi ckey, Mark Hayden, Kenneth Birman, and Robert Constable. Building reliable, high-performance communication systems from compone nts. Operating Systems Review, 34(5):80--92, December 1999. Presented at the 17 th ACM Symposium on Operating Systems Principles (SOSP'99).

....Thus, the use of the component based software development paradigm for constructing and tailoring embedded real time systems has premise. We elaborate this in more detail by listing the major benefits of using component based software engineering for development of embedded real time systems [106, 64]: D Rapid development and deployment of the system. Component based software development aims to reduce the amount of new code that must be written each time new application is being developed. Many software components, if properly designed and verified, can be reused in different embedded ....

....and hardware platforms. D Configurable systems. An architecture of a configurable system allows new components to be developed and integrated into the system. Components in such systems are true building parts of the system. A variety of configurable systems exists, e.g. VEST [102] Ensemble [64], the approach to system development with real time components introduced by Isovi c [49] and systems based on the port based object (PBO) model [105] 3.3.2 Extensible systems SPIN SPIN [12, 85] is an extensible operating system that allows applications to define customized services. Hence, ....

[Article contains additional citation context not shown here]

X. Liu, C. Kreitz, R. Renesse, J.Hickey, M. Hayden, K. Birman, and R. Constable. Building reliable, high-performance communication systems from components. In Proceedings of the 17th ACM Symposium on Operating Systems Principles (SOSP), volume 34, pages 80--92, December 1999. Published as Operating Systems Review.

No context found.

X. Liu, C. Kreitz, R. van Renesse, J. Hickey, M. Hayden, K. Birman, and R. Constable. Building Reliable, HighPerformance Communication Systems from Components. In Symposium on Operating Systems Principles, 1999.

No context found.

Xiaoming Liu, Christoph Kreitz, Robbert van Renesse, Jason J. Hickey, Mark Hayden, Kenneth Birman, and Robert Constable. Building reliable, high-performance communication systems from components. In 17 ACM Symposium on Operating Systems Principles (SOSP'99), volume 33(5) of Operating Systems Review, pages 80--92. ACM Press, December 1999.

No context found.

X. liu, C. Kreitz, R. van Renesse, J. Hickey, M. Hayden, K. Birman, and R. Constable. Building reliable, highperformance communication systems from components. In Proc. of 17th ACM Symposium on Operating Systems Principles (SOSP'99), Kiawah Island, SC, Dec. 1999.

No context found.

LIU, X., KREITZ, C., VAN RENESSE, R., HICKEY, J., HAYDEN, M., BIRMAN, K., AND CONSTABLE, R. Building reliable, high-performance communication systems from components. In Proc. of the 17 ACM SOSP (Charleston, USA, Dec 1999), pp. 80--92.

No context found.

X. Liu, et al., \Building Reliable, High-Performance Communication Systems from Components." Operating Systems Review, 34(5), pp. 80-92, Dec. 1999.

No context found.

Xiaoming Liu, Christoph Kreitz, Robbert van Renesse, Jason Hickey, Mark Hayden, Kenneth Birman, and Robert Constable. Building reliable, high-performance communication systems from components. In 17 ACM Symposium on Operating Systems Principles (SOSP'99), volume 34 of Operating Systems Review, pages 80--92, 1999.

No context found.

Xiaoming Liu, Christoph Kreitz, Robbert van Renesse, Jason Hickey, Mark Hayden, Kenneth P. Birman, and Robert L. Constable. Building reliable, high-performance communication systems from components. In Proceedings of the Seventeenth ACM Symposium on Operating Systems Principles, pages 80--92, Kiawah Island, SC, U.S.A., December 1999. ACM SIGOPS. 145.

First 50 documents

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