Toby Lehman | University of Wisconsin-Madison (original) (raw)

Papers by Toby Lehman

designed and implemented a memory-resident stor-age component that co-exists with Star-burst’s di... more designed and implemented a memory-resident stor-age component that co-exists with Star-burst’s disk-oriented storage component. The two storage compo-nents share the same common services, such as query optimization, transaction management, etc. However, the memory-resident storage component is faster than the disk-oriented storage component and hence needs faster run-time services. This paper examines two run-time services, the lock manager and the latch mecha-nism, and investigates possible cost-cutting measures. We propose the use of of a single latch for protecting a table, all of its indexes, and all of its related lock in-formation, in order to reduce storage component latch costs. VVe then show that although a table-level latch is a large granule latch, it does not significantly re-strict concurrency. VVe also examine traditional lock manager design and suggest a different design that is appropriate for memory-resident storage components. The new design exploits direct address...

Ibm Systems Journal, 1998

... Gordon E. Moore, co-founder of Intel Corporation, postulated that the capacity and capa-bilit... more ... Gordon E. Moore, co-founder of Intel Corporation, postulated that the capacity and capa-bility of computers would approximately double ev-ery 18 months. This 1965 prediction still holds today for most types of computers. ... There is an interesting side effect to Moore&amp;amp;amp;#x27;s Law. ...

IECON'99. Conference Proceedings. 25th Annual Conference of the IEEE Industrial Electronics Society (Cat. No.99CH37029)

TSpaces is a global communication middleware component that incorporates database features, such ... more TSpaces is a global communication middleware component that incorporates database features, such as transactions, persistent dates, flexible queries and XML support. TSpaces is an excellent tool for building distributed applications, since it provides an asynchronous and anonymous link between multiple clients or services. The communication link provided by TSpaces gives application builders the advantage of ignoring some of the harder aspects of multi-client synchronization, such as tracking names (and addresses) of all active clients, communication line status, and conversation status. For many applications, the loose synchronization provided by TSpaces works extremely well. However, for distributed real-time applications the loose synchronization model poses some problems. For these types of application, the TSpaces middleware must not introduce any additional unbounded latency in the interaction between network components. We describe the changes that would be needed in the current Java-based implementation of TSpaces to provide support for real-time applications, such as deterministic memory management, handling priority requests, bounded execution paths for all operations and dealing with overload situations

This paper describes the TSpaces approach to integrated control of heterogeneous devices in heter... more This paper describes the TSpaces approach to integrated control of heterogeneous devices in heterogeneous control networks. The TSpaces system is an application middleware that provides a uniform layer of indirection between components. The system provides gatewaying between legacy network protocols and IP, asynchronous group communication, a shared data schema, component rendezvous and event composition. The authors describe each of these

Journal of Revenue and Pricing Management, 2010

In over 40 years with IBM, he has managed projects including theory of programming, semantic info... more In over 40 years with IBM, he has managed projects including theory of programming, semantic information processing and distributed computing and information processing. An author/inventor of numerous publications and patents, Strong has worked in the research areas of access methods, machine architectures, communication and synchronization, fault tolerant distributed algorithms, systems management, knowledge management, and service delivery. He was a member of the project that produced the world's first Relational Database Management System. Isaac Councill works on improving search quality at Google, and received a PhD in Information Sciences and Technology from Penn State University, where he specialized in applied machine learning and information retrieval. While at IBM Research, he focused on developing tools to support the complex domain of costing and pricing global technology services.

IEEE Transactions on Software Engineering, 1984

We identify the roles played by design versions and alternation about the design of complicated "... more We identify the roles played by design versions and alternation about the design of complicated "engineered" artifacts, tives in an engineering database. The obvious way to implement versions e.g., large software systems, multiauthor documents, and inis to maintain each in a separate collection of files. Because several ver-tegrated circuits. Although the structures developed here are sions must be kept on line in a design environment, the approach leads to of use in any of these domains, we will concentrate on intelarge disk requirements. We develop B-tree-based storage structures to encode versions as "negative" differential files. Our objective is to keep grated circuits. Artifacts as complicated as these are created the disk requirements small. We discuss the effect of enormous amounts by teams of designers simultaneously working on different of cheap archival storage (write-once optical digital disks) on the pro-pieces of a design. In addition, a design may be specified in posed structures. We have implemented versions in the Wisconsin stor-several different design representations. For example, an inage system (WiSS), an experimental database component developed at tegrated circuit design is viewed simultaneously as a geometric the University of Wisconsin-Madison.tgae ici eini iwe iutnosya emtl layout, a transistor network, or a logic circuit. Besides controled sharing and multiple design representations, the environprogress version. When a design is ready for release, it is first Manuscript received August 31, 1982; revised February 4, 1983. made effective for testing and internal distribution. Effective This work was supported by the Wisconsin Alumni Research Fund, by versions cannot be updated without once again becoming inthe National Science Foundation under Grant MCS-8201860, and by a progress. Effective versions become released when the designed grant from the IBM Corporation. R. H. Katz was with the Department of Computer Sciences, Univer-object is sent into the field. Released versions can be archived sity of Wisconsin, Madison, WI 53706. He is now with Department of and later restored when immediate access is necessary. Electrical Engineering and Computer Sciences, University of California, We are concerned with how to use existing database tech-Berkeley, CA 94720.

IBM Systems Journal, 1998

With the creation of computer networks in the 1970s came the birth of distributed network applica... more With the creation of computer networks in the 1970s came the birth of distributed network applications. Since then, there have been many applications that spanned multiple machines, but in the last 20 years no one created a serviceable network middleware package for developing highly effective distributed applications, that is, until now. This paper describes the design and architecture of T Spaces, a project at the IBM Almaden Research Center that fills the network middleware void. T Spaces embodies the three main characteristics of a useful mechanism for network programs, namely, data management, computation, and communication. Since it has the potential to connect any program to any other program on a computing network, T Spaces is an ideal platform on which to build a global computing services platform where any program or system service is available to any other program or service. In addition, its small footprint and Java™ implementation make T Spaces an ideal platform for writing distributed applications for embedded and palm-top computers, thus forging a needed gateway from the emerging embedded and palm-top computers to established desktop and server computers.

IBM Systems Journal, 1999

Computer Networks, 2001

Our world is becoming increasingly heterogeneous, decentralized and distributed, but the software... more Our world is becoming increasingly heterogeneous, decentralized and distributed, but the software that is supposed to work in this world, usually, is not. TSpaces is a communication package whose purpose is to alleviate the problems of hooking together disparate distributed systems. TSpaces is a global communication middleware component that incorporates database features, such as transactions, persistent data,¯exible queries and XML support. TSpaces is an excellent tool for building distributed applications, since it provides an asynchronous and anonymous link between multiple clients or services. The communication link provided by TSpaces gives application builders the advantage of ignoring some of the harder aspects of multi-client synchronization, such as tracking names (and addresses) of all active clients, communication line status, and conversation status. For many dierent types of applications, the loose synchronization provided by TSpaces works extremely well. This paper relates our experiences in building distributed systems with TSpaces as the central communication component.

Millions of small heterogeneous computers are poised to spread into the infrastructure of our soc... more Millions of small heterogeneous computers are poised to spread into the infrastructure of our society. Though mostly inconspicuous today, disguised as nothing more than PIM (personal information management) computers, these tiny processors will eventually pervade most aspects of civilized life. The one thing holding them back from being everyone's portal to the new electronic society and the access point to an in nite store of information is the lack of a high-quality logical link to the world's network backbone. Enter T Spaces, a network middlewarepackage for the new age of ubiquitous computing. T Spaces is a tuplespace-based network communication bu er with database capabilities that enables communication between applications and devices in a network of heterogeneous computers and operating systems. With T Spaces, it is possible to connect all computers together, which leads the way towards an in nitely large cluster of cooperating machines. In this paper we describe the...

Proceedings of the 32nd Annual Hawaii International Conference on Systems Sciences. 1999. HICSS-32. Abstracts and CD-ROM of Full Papers

Page 1. T Spaces: The Next Wave Tobin J. Lehman Stephen W. McLaughry Peter Wycko IBM Almaden Rese... more Page 1. T Spaces: The Next Wave Tobin J. Lehman Stephen W. McLaughry Peter Wycko IBM Almaden Research Center University of Oregon Bellcore toby@almaden.ibm.com stephen@cs.uoregon.edu wycko @bellcore.com Abstract ...

Web services allow authorized entities (including individuals, corporations, and automated agents... more Web services allow authorized entities (including individuals, corporations, and automated agents) to employ software components created by other parties scattered across the globe. In support of this powerful model of interaction, we have designed and implemented an infrastructure and a set of tools to simplify the development and management of Web services. This infrastructure and these tools form the TSpaces Services Suite (TSSuite). This paper shows how TSSuite supports the Web services model, and it describes the design and functionality of the main TSSuite components.

As part of the Starburst extensible database project developed at the IBM Almaden Research Center... more As part of the Starburst extensible database project developed at the IBM Almaden Research Center, we designed and implemented a memory-resident storage component that co-exists with Star-burst’s diskoriented storage component. The two storage components share the same common services, such as query optimization, transaction management, etc. However, the memory-resident storage component is faster than the disk-oriented storage component and hence needs faster run-time services. This paper examines two runtime services, the lock manager and the latch mechanism, and investigates possible cost-cutting measures. We propose the use of of a single latch for protecting a table, all of its indexes, and all of its related lock information, in order to reduce storage component latch costs. VVe then show that although a table-level latch is a large granule latch, it does not significantly restrict concurrency. VVe also examine traditional lock manager design and suggest a different design tha...

PURPOSE: To support the storage, access and maintenance of an extremely long field by permitting ... more PURPOSE: To support the storage, access and maintenance of an extremely long field by permitting a database table adding a long field value to add two files in connection with it. CONSTITUTION: A data file 16 consisting of a database record having plural related and long field descriptors 22 is generated and a long field file 18 consisting of a long field record having related distribution page data is generated. Then, a cancelling log is generated in functional response to change as against the database and the long field descriptor is restored in the corresponding database record and distribution page data in response to the cancelling log. Thus, the long field in the database record is supported without supporting records over plural pages, copying the value of the long field plural times and storing an intermediate buffer, etc.