Attiya H., Herlihy M.P. and Rachman O., Atomic Snapshots Using Lattice Agreement, Distributed Computing, 8(3):121--132, 1995.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....started and the collect nished. Curiously, the trivial implementation is the one used in almost all of the many asynchronous shared memory algorithms based on collects, including algorithms for consensus, snapshots, coin ipping, bounded round numbers, timestamps, and multi writer registers [1, 2, 5, 6, 8, 9, 12, 13, 15, 24 26, 29 31, 33, 35, 36, 38, 40 42, 52]. Noteworthy exceptions are [49, 50] which present interesting collect algorithms that do not follow the pattern of the trivial algorithm, but which depend on making strong assumptions about the schedule. Part of the reason for the popularity of this approach may be that the trivial algorithm ....

.... [11] to prove low competitive throughput for an algorithm that improves on the algorithm of [3] We show in Section 6 that relative competitiveness, combined with a throughput competitive collect algorithm, does in fact give throughput competitive solutions to problems such as atomic snapshot [2, 5, 9, 13, 15]and bounded round numbers [31] and argue that most algorithms that use collects can be shown to be throughputcompetitive using similar techniques. Finally, in Section 7 we discuss some related approaches to analyzing distributed algorithms and consider what questions remain open. 8 2 Model We ....

[Article contains additional citation context not shown here]

Hagit Attiya, Maurice Herlihy, and Ophir Rachman. Atomic snapshots using lattice agreement. Distributed Computing, 8(3):121-132, 1995.

....of a radar tracking system, where multiple sensors generate updates concurrently with multiple requests for consistent system states. The design of asymptotically efficient implementations of an atomic snapshot memory has been the subject of extensive and highly creative research in recent years [3, 6, 8, 11, 13, 18, 21, 30]. An atomic snapshot memory is an abstract data type equivalent to a memory partitioned into n segments, one for each processor. There are two types of operations on the object, a scan and an update. In an update operation, a processor writes the contents of its associated segment, while in a ....

....with only O(1) update complexity (in fact, at most four operations) O(n) scan complexity, and O(n 2 ) space complexity. Though one might think that the use of strong primitives like Load linked Store conditional would allow us to readily modify the elegant snapshot algorithms in the literature [3, 6, 8, 11, 13, 18, 21] to achieve similar complexity, it turns out that this is not the case (see the summary in Table 1) These multi scanner snapshot protocols have an algorithmic structure in which each updater and or scanner collects a view of memory in its register, and then processors try to agree which of the ....

[Article contains additional citation context not shown here]

Attiya, H., Herlihy, M., and Rachman, O. Atomic snapshots using lattice agreement. Distributed Computing, 8(3), pages 121--132.

.... 25] combinatorial properties and impossibility results [1, 9, 10, 11, 12, 24, 29] linearizability properties [22, 27, 28] analysis of their performance by both theoretical and experimental means [2, 4, 8, 13, 15, 16, 18, 20, 21, 22, 23, 30, 31, 32] and applications to solving decision problems [5]. The principal motivation for introducing counting and smoothing networks has been to implement shared counters in a way to ensure scalable performance as the number of concurrent processors n, called concurrency, that may simultaneously access the counter grows large. Indeed, overwhelming ....

H. Attiya, M. Herlihy and O. Rachman, "Atomic Snapshots Using Lattice Agreement," Distributed Computing, Vol. 8, pp. 121--132, 1995.

....which are comparable in the lattice. The lattice agreement problem is closely related to the problem of implementing atomic snapshot objects [1] A wait free lattice agreement algorithm can be turned into wait free implementation of atomic snapshots, with O(n) additional read write operations [5]. This paper presents the following algorithms: An adaptive O(k log k) algorithm for lattice agreement. An adaptive O(k log k) algorithm for one time (6k Gamma 1) renaming. A fast O(K log K) algorithm for one time (2k Gamma 1) renaming. The step complexity of the best existing fast ....

....best existing algorithm which reduces the size of name space from N to 2k Gamma 1 has step complexity O(Nk 2 ) 7] The linear renaming algorithm can be modified into a linear lattice agreement algorithm which uses only dynamic single reader single writer registers. Using the transformation of [5], this gives an O(n) implementation of atomic snapshot using dynamic multi reader single writer registers. The best previously known algorithms for atomic snapshot are an O(n log n) implementation using multi reader singlewriter registers [6] and an O(n) implementation using multi reader ....

[Article contains additional citation context not shown here]

H. Attiya, M. Herlihy, and O. Rachman. Atomic snapshots using lattice agreement. Distributed Computing, 8(3):121--132, 1995.

....Since the rst appearance of the preliminary versions of this paper [7, 25] there have been many advances in the study of wait free objects built from atomic registers. In particular, there has been considerable improvement in algorithms for atomic snapshots. The lattice agreement technique [8], where processes agree on a chain in a lattice, is closely related to the semilattice construction we use in Section 6. By allowing processes to obtain values spread throughout the lattice instead of pushing all processes toward the top, lattice agreement allows for faster snapshot protocols such ....

Hagit Attiya, Maurice Herlihy, and Ophir Rachman. Atomic snapshots using lattice agreement. Distributed Computing, 8(3):121-132, 1995.

....problem [1] extends the collect problem by requiring views to look instantaneous. Instead of separate update and store operations, we provide a combined upscan operation, which updates a new value and atomically collects a view. The returned views should satisfy the following conditions (cf. [14]) Validity: If an upscan operation op returns a view V , and precedes an upscan operation op 0 , then V does not include the value written by op 0 . 2 Self inclusion: The view returned by the th upscan operation of p j includes the th value written by p j . Comparability: If V 1 and V 2 ....

....O(n log n) steps [15] This transformation is not trivial since in the non adaptive algorithm, processes descend down a binary tree of depth O(log n) see below) thus, the number of stores and collect depends on n. We describe the one shot algorithm; it can be made long lived using techniques of [14, 15, 20]. The non adaptive algorithm uses a complete binary tree of depth log n, whose vertices are labeled as a search tree, in which all values are stored in the leaves: The leaves are labeled 2 Typically, this condition trivially holds and we do not prove it below. 9 Algorithm 2 The classifier ....

H. Attiya, M. Herlihy, and O. Rachman. Atomic snapshots using lattice agreement. Dist. Comp., 8(3):121--132, 1995.

....of processes in each set, and allows them to employ a range independent (2k Gamma 1) Adaptive Lattice Agreement and Renaming 3 renaming algorithm designed for this bound. Different sets use disjoint name spaces; no coordination between the sets is required. In the lattice agreement problem [15], processes obtain comparable (by containment) subsets of the set of active processes. A wait free lattice agreement algorithm can be turned into a wait free implementation of an atomic snapshot object, with O(n) additional read write operations [15] Atomic snapshot objects allow processes to get ....

....is required. In the lattice agreement problem [15] processes obtain comparable (by containment) subsets of the set of active processes. A wait free lattice agreement algorithm can be turned into a wait free implementation of an atomic snapshot object, with O(n) additional read write operations [15]. Atomic snapshot objects allow processes to get instantaneous global views ( snapshots ) of the shared memory and thus, they simplify the design of wait free algorithms. The step complexity of our adaptive algorithm for lattice agreement is O(k log k) In this algorithm, processes first obtain ....

[Article contains additional citation context not shown here]

H. Attiya, M. Herlihy, and O. Rachman, Atomic snapshots using lattice agreement, Dist. Comp., 8 (1995), pp. 121--132.

....extension of the atomic snapshot problem; it supports a combined im upscan operation, which updates a new value and returns a view. In this case, the atomic snapshot properties are guaranteed if the views returned in some execution of im upscan operations satisfy the the following conditions (cf. [13]) Validity: If an im upscan operation op returns a view V , and precedes an im upscan operation op 0 , then V does not include the value written by op 0 . 1 Self inclusion: The view returned by the th im upscan operation of p j includes the th value written by p j . Comparability: If V ....

H. Attiya, M. Herlihy, and O. Rachman. Atomic snapshots using lattice agreement. Distributed Computing, 8(3):121--132, 1995.

....paper presents adaptive wait free algorithms for lattice agreement and renaming, using only read and write operations. Along the way, we improve the step complexity of non adaptive algorithms for renaming. Figure 1 depicts the algorithms presented in this paper. In the lattice agreement problem [11], each process should obtain a subset of the active processes and the subsets have to be comparable, i.e. ordered by the containment relation. A wait free lattice agreement algorithm can be turned into a wait free implementation of an atomic snapshot object, with O(n) additional read write ....

....process should obtain a subset of the active processes and the subsets have to be comparable, i.e. ordered by the containment relation. A wait free lattice agreement algorithm can be turned into a wait free implementation of an atomic snapshot object, with O(n) additional read write operations [11]. An atomic snapshot object allows processes to get an instantaneous global view ( snapshot ) of the shared memory, and is useful for simplifying the design of wait free algorithms. In the M renaming problem [8] processes are required to choose a distinct name in a range of size M . This paper ....

[Article contains additional citation context not shown here]

H. Attiya, M. Herlihy, and O. Rachman. Atomic snapshots using lattice agreement. Distributed Computing, 8(3):121--132, 1995.

....fast wait free algorithms for lattice agreement and renaming. To the best of our knowledge, these are the first fast wait free algorithms which use only read write operations. Along the way, we improve the step complexity of non fast algorithms for these problems. In the lattice agreement problem [6], each process must obtain a subset of the active processes and the subsets have to be comparable, i.e. ordered by the containment relation. A wait free lattice agreement algorithm can be turned into a wait free implementation of an atomic snapshot object, with O(n) additional read write ....

....process must obtain a subset of the active processes and the subsets have to be comparable, i.e. ordered by the containment relation. A wait free lattice agreement algorithm can be turned into a wait free implementation of an atomic snapshot object, with O(n) additional read write operations [6]. An atomic snapshot object allows processes to get an instantaneous global view ( snapshot ) of the shared memory, and is useful for simplifying the design of wait free algorithms. In the M renaming problem [4] processes are required to choose a distinct name in the range f0 : M Gamma 1g, ....

[Article contains additional citation context not shown here]

Hagit Attiya, Maurice Herlihy, and Ophir Rachman. Atomic snapshots using lattice agreement. Distributed Computing, 8(3):121--132, 1995.

No context found.

Attiya H., Herlihy M.P. and Rachman O., Atomic Snapshots Using Lattice Agreement, Distributed Computing, 8(3):121--132, 1995.

No context found.

Attiya H., Herlihy M.P., Rachman O., Atomic Snapshots Using Lattice Agreement, Distributed Computing 8(3), 121--132 (1995)

No context found.

Attiya H., Herlihy M.P. and Rachman O., Atomic Snapshots Using Lattice Agreement, Distributed Computing, 8(3):121--132, 1995.

No context found.

Attiya H., Herlihy M.P. and Rachman O., Atomic Snapshots Using Lattice Agreement, Distributed Computing,8[puting[ 1995.

No context found.

H. Attiya, M. Herlihy, and O. Rachman, Atomic snapshots using lattice agreement, Distributed Computing, 8 (1995) 121--132.

No context found.

Attiya H., Herlihy M.P. and Rachman O. Atomic Snapshots Using Lattice Agreement. Distributed Computing, 8(3):121-132, 1995.

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