Latest Articles from JUCS - Journal of Universal Computer Science Latest 5 Articles from JUCS - Journal of Universal Computer Science https://lib.jucs.org/ Thu, 28 Mar 2024 22:42:33 +0200 Pensoft FeedCreator https://lib.jucs.org/i/logo.jpg Latest Articles from JUCS - Journal of Universal Computer Science https://lib.jucs.org/ Scalable Distributed Metadata Server Based on Nonblocking Transactions https://lib.jucs.org/article/23991/ JUCS - Journal of Universal Computer Science 26(1): 89-106

DOI: 10.3897/jucs.2020.006

Authors: Kohei Hiraga, Osamu Tatebe, Hideyuki Kawashima

Abstract: Metadata performance scalability is critically important in high-performance computing when accessing many small files from millions of clients. This paper proposes a design of a scalable distributed metadata server, PPMDS, for parallel file systems using multiple key-value servers. In PPMDS, hierarchical namespace of a file system is efficiently managed by multiple servers. Multiple entries can be atomically updated using a nonblocking distributed transaction based on an algorithm of dynamic software transactional memory. This paper also proposes optimizations to further improve the metadata performance by introducing a server-side transaction processing, multiple readers, and a shared lock mode, which reduce the number of remote procedure calls and prevent unnecessary blocking. Performance evaluation shows the scalable performance up to 3 servers, and achieves 62,000 operations per second, which is 2.58x performance improvement compared to a single metadata performance.

HTML

XML

PDF

]]>
Research Article Tue, 28 Jan 2020 00:00:00 +0200
Cyclic Reference Counting with Permanent Objects https://lib.jucs.org/article/28811/ JUCS - Journal of Universal Computer Science 13(6): 830-838

DOI: 10.3217/jucs-013-06-0830

Authors: Rafael Lins, Francisco Heron de Carvalho Júnior, Zanoni Lins

Abstract: Reference Counting is the memory management technique of most widespread use today. Very often applications handle objects that are either permanent or get tenured. This paper uses this information to make cyclic reference counting more efficient.

HTML

XML

PDF

]]>
Research Article Thu, 28 Jun 2007 00:00:00 +0300
A Formal Semantics for Finalizers https://lib.jucs.org/article/28441/ JUCS - Journal of Universal Computer Science 11(7): 1198-1214

DOI: 10.3217/jucs-011-07-1198

Authors: Marcus Leal, Roberto Ierusalimschy

Abstract: Automatic finalization is a common but inherently complex language facility that makes the garbage collection process semantically visible to client programs. With finalizers, memory management becomes more flexible, and garbage collectors can be used to recycle other resources in addition to memory. Formal language models usually ignore garbage collection, and therefore are unable to properly describe finalization. In this paper we use an operational approach to develop a new abstract model that explicitly represents memory management actions in a garbage­collected programming language based on the λ­calculus. We formally state and prove several important properties related to memory management, and employ the model to describe and explore a semantics for finalizers.

HTML

XML

PDF

]]>
Research Article Thu, 28 Jul 2005 00:00:00 +0300
Lazy Cyclic Reference Counting https://lib.jucs.org/article/28084/ JUCS - Journal of Universal Computer Science 9(8): 813-828

DOI: 10.3217/jucs-009-08-0813

Authors: Rafael Lins

Abstract: Reference counting is a widely employed memory management technique, in which garbage collection operations are interleaved with computation. Standard reference counting has the major drawback of being unable to handle cyclic structures. This paper presents an important optimisation to a recently published algorithm for cyclic reference counting. Proofs of the correctness of the original and lazy algorithms are provided, together with performance figures.

HTML

XML

PDF

]]>
Research Article Thu, 28 Aug 2003 00:00:00 +0300
Distributed Caching in Networked File Systems https://lib.jucs.org/article/27131/ JUCS - Journal of Universal Computer Science 1(6): 399-409

DOI: 10.3217/jucs-001-06-0399

Authors: Artur Klauser, Reinhard Posch

Abstract: Changing relative performance of processors, networks, and disks makes it necessary to reconsider algorithms using these three resources. As networks get faster and less congested topologies emerge, it becomes important to use network resources more aggressively to obtain good performance. Substitution of local disk accesses by accesses to remote memory can lead to better balanced resource usage and thus to faster systems. In this work we address the issue of file caching in a networked file system configuration. Distributed block-level in-memory caches are considered. We show that carefully constructed distributed concepts can lead to lower server load and better overall system performance than centralized concepts. Oversimplification, although aimed at gaining performance for single components, may deteriorate overall performance as a result of unbalanced resource usage.

HTML

XML

PDF

]]>
Research Article Wed, 28 Jun 1995 00:00:00 +0300