David R. Cheriton. V Kernel: a Software Base for Distributed Systems. IEEE Software, 1(2):19-38, 40-42, April 1984.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....for file updates and degree 1 consistency on directories. presents the overall architecture of transaction management in QuickSilver and describes the toolkit in more detail. Transactional IPC. IPC in QuickSilver follows the request response paradigm. It is similar to IPC in the V system [4], except that QuickSilver IPC requests may be made asynchronously. A unique feature of QuickSilver IPC is that it is transactional; that is, all interprocess communication must be done on behalf of some trans action. There is no escape from this aspect of transactions in QuickSilver. Every IPC ....

David R. Chefiron. The V kernel: a software base for distributed systems. IEEE Software, 1(2):19- 42, April 1984.

....amounts of data (currently 8192 bytes) The small pages ( short pages ) are useful for synchronization and storing small objects; the large pages are useful for bulk data transfer. These sizes are a good match to most protocol data unit sizes encountered in networking environments[Caceres 89] Cheriton 84] Figure 3.4 gives a block overview of the MNFS address space and how clients map the MNFS memory space into their own addressable space. Each client shares the entire segment of MNFS space with all the other clients who have mapped in that segment. Mem Client Mem Client Mem Client MNFS ....

David R. Cheriton. The V Kernel: A Software Base for Distributed Systems. IEEE Software; Vol. 1, No. 2, April 1984. pp 19-42.

....cost but are useful for moving large amounts of data (currently 8192 bytes) The small pages ( short pages ) are useful for synchronization and small objects; the large pages are useful for bulk data transfer. These sizes are well matched to the sizes most commonly seen in networking environments[3][7] Locks on Pages It is possible to lock a page down in those cases where exclusive write access for a period of time is required. Typically, locks are used only for initialization. Event Driven Memory Synchronization In Event Driven Memory Synchronization (EDMS) one process can pause in the ....

D.R. Cheriton. The v kernel: A software base for distributed systems. IEEE Software, 1(2):19--42, April 1984.

....cost but are useful for moving large amounts of data (currently 8192 bytes) The small pages ( short pages ) are useful for synchronization and small objects; the large pages are useful for bulk data transfer. These sizes are well matched to the sizes most commonly seen in networking environments[2][5] 1 assuming, of course, that the SPARC implementation in question executes the instruction correctly; not all do 3 main( f int fd; int x, shortpage; int i; SETUP open file and map it in. fd = open( mnfs test , O RDWR O CREAT, 0777) map the file in let the ....

D.R. Cheriton. The v kernel: A software base for distributed systems. IEEE Software, 1(2):19--42, April 1984.

....and recovery. These primitives facilitate the implementation of faulttolerant process groups, which can be used to provide distributed services in an environment subject to non malicious crash failures. Unlike other process group approaches, such as Cheriton s host groups (RFC s 966, 988, [Cheriton]) our approach provides powerful guarantees about the behavior of the communication subsystem when process group membership is changing dynamically, for example due to process or site failures, recoveries, or migration of a process from one site to another. Our approach also addresses delivery ....

....remote procedure calls on process groups [Birman b] The approach to communication that we report here arose as part of this new version of the ISIS system. Unreliable communication primitives, such as the multicast group communication primitives proposed in RFC s 966 and 988 and in [Cheriton ], leave some uncertainty in the delivery status of a message when failures and other exceptional events occur during communication. Instead, a form of best effort delivery is provided, but with the possibility that some member of a group of processes did not receive the message if the group ....

[Article contains additional citation context not shown here]

Cheriton, D. The V Kernel: A software base for distributed systems. IEEE Software 1 12, (1984), 19-43.

....multiprocessor. Multiprocessor workstations have already appeared in the marketplace. Operating systems in which multiprocessor workstations can be efficiently used have been developed. Typical multiprocessor operating systems are Mach [18] developed at CMU, and the V distributed operating system [4] developed at Stanford university. In this paper, the design and implementation of a version of Smalltalk, called HoME (HPS on Mach Environment) 13] which can be efficiently executed on a multiprocessor are described. The multiprocessor used is the shared memory multiprocessor OMRON LUNA88K. ....

.... (MS) 14] 15] and ConcurrentSmalltalk (CST) 21] MS is an example of Smalltalk where processes (units of execution) execute in parallel like HoME, and it is a system which implements Berkeley Smalltalk [19] on the FireFly multiprocessor [17] with the V distributed operating system [4]. Berkeley Smalltalk is a type which unlike HPS interprets byte code. In MS, concurrent execution is realized by replicating the interpreter and having the same number of interpreters as the number of physical processors. The ready queue for processes is not replicated. There is only one ready ....

David R. Cheriton, The V kernel: a software base for distributed systems, IEEE Software, 1 (2), April 1984.

....large amounts of data (currently 8192 bytes) The small pages ( short pages ) are useful for synchronization and storing small objects; the large pages are useful for bulk data transfer. These sizes are well matched to the protocol data unit sizes most commonly seen in networking environments[10][7] 8 Event Driven Memory Synchronization In Event Driven Memory Synchronization (EDMS) one process can pause in the middle of a memory read (i.e. after the address has been issued but before the data cycle is complete) and will continue only after another process has taken some action. This ....

D.R. Cheriton. The v kernel: A software base for distributed systems. IEEE Software, 1(2):19-- 42, April 1984.

....with a mobile software component has been tackled before by research on process migration. At their core, the techniques adopted in this area are by and large the same as we have discussed so far. For example, Sprite [8] and DEMOS MP [14] both use a forwarding strategy, while the V kernel [7] describes a broadcast protocol to rediscover the location of a lost process. Nevertheless, the model underlying process migration assumes tight control of the system over the movement of processes, which usually does not occur very frequently or over long distances. Under these constraints, ....

D.R. Cheriton. The V Kernel: A Software Base for Distributed Systems. IEEE Software, 1(2):19--42, April 1984.

....where it originated. Here we limit the discussion to the scenario where all the participating machines are running (possibly different versions of) Unix. This is because specially designed distributed operating systems typically use different mechanisms that are tailor made for them [30] 22] 21] [19] [20] 23] The problem is also slightly simpler for such operating systems since they have been designed from scratch keeping the goals of distribution in mind. Thus they typically incorporate mechanisms that allow all resources to be shared in a transparent and uniform manner. Unix, on the other ....

David R. Cheriton, The V kernel: A software base for distributed systems, IEEE Software 1(2), April 1984.

....of LAN technology and workstations. The sheer bulk of this software is an indication that it may be overwhelming our ability to create reliable, efficient code: for example, 30 of the 4.3BSD Unix [8, 21] kernel source, 25 of the TOPS 20 [10] Version 6. 1) kernel source, and 32 of the V system [4] kernel source are devoted to networking. Development of network software is slow and seldom yields finished systems; debugging may continue long after the software is put into operation. Continual debugging of production code results not only from deficiencies in the original code, but also from ....

David R. Cheriton. The V Kernel: A software base for distributed systems. IEEE Software 1(2):19-42, April, 1984.

....cost but are useful for moving large amounts of data (currently 8192 bytes) The small pages ( short pages ) are useful for synchronization and small objects; the large pages are useful for bulk data transfer. These sizes are well matched to the sizes most commonly seen in networking environments[12][17] Event Driven Memory Synchronization Event Driven Memory Synchronization (EDMS) allows a process to pause waiting for another process to issue a network refresh. Thus the resumption of execution of a process is a direct consequence of an action taken by a process on a remote machine. This ....

D.R. Cheriton. The v kernel: A software base for distributed systems. IEEE Software, 1(2):19--42, April 1984.

....have to take place for each packet. FLIP is implemented using stop and wait at the packet level i.e. Idle RQ. At the transaction layer, timeouts and retransmissions are used to provide reliable transactions. 2. 4 V s VMTP The Versatile Message Transaction Protocol was designed for the V Kernel [7] at Stanford University. VMTP is a transport protocol designed to support the transaction model of communication, as exemplified by the Remote Procedure Call . The basic protocol is a request response protocol. A message is sent as one or more packet groups. A packet group consists of up to 16k ....

CHERITON, D. R. (1984). The V Kernel: A Software Base for Distributed Systems. IEEE Software, April 1984, 19.

....of the ARCADE project is to provide an environment in which tasks can cooperate in a uniform and effective manner even if they reside on machines with radically different hardware architectures. This objective distinguishes ARCADE from other distributed systems, such as Mach [1, 21] V kernel [8], and CHORUS [2] While these systems also adopt the minimal kernel approach, they are not overly concerned with the problems of heterogeneity. In keeping with the minimalist philosophy, ARCADE defines only two abstractions: active tasks and passive data units. For the purposes of this discussion, ....

Cheriton, D., The V Kernel: A Software Base for Distributed Systems, IEEE Software, April, 1984, pp. 19-42.

....6 summarizes our results. 2 Interprocess Communication Interprocess communication is central to modern operating system structure and performance. Operating systems have evolved from monolithic, centralized kernels to a more decentralized structure [Baskett et al. 75, Rashid Robertson 81, Cheriton 84, Jones Rashid 86] The reasons for this evolution are two fold. First, by structuring the operating system as independent address spaces communicating through messages, modularity, fault tolerance, and extensibility are enhanced. Second, using messages rather than shared memory as the principal ....

D. R. Cheriton. The V kernel: A software base for distributed systems. IEEE Software, 1(2):19--42, April 1984.

....process model to construct family members was taken from Thoth [3] Peace is the evolutionary successor of Ax. The design strongly followed the idea of Moose and relied on an extremely efficient microkernel as minimal basis for the operating system family. Important influences came from V [4], Amoeba [11] and Mach [21] Compared to its ancestors , the Peace microkernel was of significantly less complexity and higher performance. 1982) 1984) 1986) 1990) Choices Mach UNIX MOOSE AX PEACE objective PEACE FAMOS DAS operating system family Amoeba QNX V Thoth Figure 1: ....

D. R. Cheriton. The V Kernel: A Software Base for Distributed Systems. IEEE Software, 1(2):19--43, 1984.

....for non blocking message passing primitives. If the overhead of VPs is low enough, this approach becomes attractive. However, OD at the granularity of OS processes leads to excessive overhead. Attempts to address the high cost of OS processes have introduced a new OS abstraction called the thread [12, 9, 6, 19]. Like processes, threads have a register context and a stack. However, unlike processes, threads do not have their own private address space. Consequently, thread switches can be cheaper than process switches because they need not involve virtual memory context switches. Similarly, local ....

....work There is a wide body of work that addresses finer virtual processor granularity than that of OS processes. These approaches can be broadly classified as OS thread based, user level thread based, language based, and object based approaches. Operating systems such as Chorus [12] Mach [9] V [6], and Solaris [19] provide OS threads that can be used to reduce the cost of OD. Context switches between OS threads of the same process do not require the switching of the virtual memory context. Consequently, thread context switch is generally an order of magnitude or so faster than process ....

David R. Cheriton. The V kernel: A software base for distributed systems. IEEE Software, 1(2):19-- 42, April 1984.

....Service DNS, electronic mail services, and others. Distributed services are programmed using various distributed programming models. Perhaps the most common example of a distributed programming model is remote procedure call (RPC) Variations of RPC include message passing as seen in the V system [16] and CSP [33] Another example of a distributed programming model is distributed shared memory (DSM) such as Ivy [41] Distributed programming models are implemented on top of a lower level abstraction referred to, in this thesis, as a network access model. An example of a network access model is ....

....Further, the SHRIMP design does not focus on issues of separating control and data transfer in distributed systems. 8. 9 Channel Model, V System, Network Objects The distributed system structure we propose is loosely related to the Channel Model [31] Network Objects [9] and other systems, like V [16], that use RPC for small data and a separate bulk data transport mechanism. Unlike most of these systems, in our model, there is no explicit activity or thread of control at the destination process to handle an incoming stream of data. Also, no specific request is required by the receiver to ....

David R. Cheriton. The V kernel: A software base for distributed systems. IEEE Software, 1(2):19--42, April 1984. 105

No context found.

D.R. Cheriton. The V Kernel: A Software Base for Distributed Systems. IEEE Software 1(2), April, 1984. 31

No context found.

David R. Cheriton. V Kernel: a Software Base for Distributed Systems. IEEE Software, 1(2):19-38, 40-42, April 1984.

No context found.

David R. Cheriton. The V Kernel: A Software Base for Distributed Systems. IEEE Transactions on Software Engineering, 1(2):19-42, April 1984.

No context found.

David R. Cheriton. The V kernel: A software base for distributed systems. IEEE Software, 1(2):19-42, April 1984.

No context found.

Cheriton, D. R. The V kernel: A software base for distributed systems. IEEE Software 1 (2): 19-42, 1984.

No context found.

D. R. Cheriton. The v kernel: A software base for distributed systems. IEEE Software, 1(2):19--42, 1984.

No context found.

David R. Cheriton. The V kernel: A software base for distributed systems. IEEE Software, 1(2):19--42, April 1984.

No context found.

D. Cheriton. The V kernel: A software base for distributed systems. IEEE Software 1(2):19--42, April 1984.

First 50 documents  Next 50

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