A distributed shared memory (DSM) system allows shared memory parallel programs to be executed on distributed memory multiprocessors. The challenge in building a DSM system is to achieve good performance over a wide range of shared memory programs without requiring extensive modifications to the source code. The performance challenge translates into reducing the amount of communication performed by the DSM system to that performed by an equivalent message passing program. This thesis describes four novel techniques for reducing the communication overhead of DSM, including: (i) the use of software release consistency, (ii) support for multiple consistency protocols, (iii) a multiple writer protocol, and (iv) an update timeout mechanism. Release consistency allows modifications of shared data to be handled via a delayed update queue, which masks network latencies. Providing multiple cons...

The memory consistency model (or memory model) of a shared-memory multiprocessor system influences both the performance and the programmability of the system. The simplest and most intuitive model for programmers, sequential consistency, restricts the use of many performance-enhancing optimizations exploited by uniprocessors. For higher performance, several alternative models have been proposed. However, many of these are hardware-centric in nature and difficult to program. Further, the multitude of many seemingly unrelated memory models inhibits portability. We use a 3P criteria of programmability, portability, and performance to assess memory models, and find current models lacking in one or more of these criteria. This thesis establishes a unifying framework for reasoning about memory models that leads to models that adequately satisfy the 3P criteria. The first contribution of this thesis is a programmer-centric methodology, called sequential consistency normal form (SCNF), for specifying memory models. This methodology is based on the observation that performance enhancing optimizations can be allowed without violating sequential consistency if the system is given some information about the program. An SCNF model is a contract between the system and the programmer, where the system guarantees both high performance and sequential consistency only if the programmer provides certain information about the program. Insufficient information gives lower performance, but incorrect information

The emerging edge services architecture promises to improve the availability and performance of web services by replicating servers at geographically distributed sites. A key challenge in such systems is data replication and consistency so that edge server code can manipulate shared data without incurring the availability and performance penalties that would be incurred by accessing a traditional centralized database. This paper explores using a distributed object architecture to build an edge service system for an e-commerce application, an online bookstore represented by the TPC-W benchmark. We take advantage of application specific semantics to design distributed objects to manage a specific subset of shared information using simple and effective consistency models. Our experimental results show that by slightly relaxing consistency within individual distributed objects, we can build an edge service system that is highly available and efficient. For example, in one experiment we find that our object-based edge server system provides a factor of five improvement in response time over a traditional centralized cluster architecture and a factor of nine improvement over an edge service system that distributes code but retains a centralized database.

We present PRACTI, a new approach for large-scale replication. PRACTI systems can replicate or cache any subset of data on any node (Partial Replication), provide a broad range of consistency guarantees (Arbitrary Consistency), and permit any node to send information to any other node (Topology Independence). A PRACTI architecture yields two significant advantages. First, by providing all three PRACTI properties, it enables better trade-offs than existing mechanisms that support at most two of the three desirable properties. The PRACTI approach thus exposes new points in the design space for replication systems. Second, the flexibility of PRACTI protocols simplifies the design of replication systems by allowing a single architecture to subsume a broad range of existing systems and to reduce development costs for new ones. To illustrate both advantages, we use our PRACTI prototype to emulate existing server replication, client-server, and object replication systems and to implement novel policies that improve performance for mobile users, web edge servers, and grid computing by as much as an order of magnitude.

To alleviate scalability problems in the Web, many researchers concentrate on how to incorporate advanced caching and replication techniques. Many solutions incorporate object-based techniques. In particular, Web resources are considered as distributed objects offering a well-defined interface. We argue that most proposals ignore two important aspects. First, there is little discussion on what kind of coherence should be provided. Proposing specific caching or replication solutions makes sense only if we know what coherence model they should implement. Second, most proposals treat all Web resources alike. Such a one-size-fits-all approach will never work in a wide-area system. We propose a solution in which Web resources are encapsulated in physically distributed shared objects. Each object should encapsulate not only state and operations, but also the policy by which its state is distributed, cached, replicated, migrated, etc.

The SHRIMP multicomputer provides virtual memorymapped communication (VMMC), which supports protected, user-level message passing, allows user programs to perform their own buffer management, and separates data transfers from control transfers so that a data transfer can be done without the intervention of the receiving node CPU. An important question is whether such a mechanism can indeed deliver all of the available hardware performance to applications which use conventional message-passing libraries. This paper

: A shared memory built on top of a distributed system constitutes a Distributed Shared Memory (DSM). If a lot of protocols implementing DSMs in various contexts have been proposed, no set of homogeneous definitions has been given for the many semantics offered by these implementations. This paper provides a suite of such definitions for atomic, sequential, causal, PRAM and a few others consistency criteria. These definitions are based on a unique framework: a parallel computation is defined as a partial order on the set of read and write operations invoked by processes, and a consistency criterion is defined as a constraint on this partial order. Such an approach provides a simple classification of consistency criteria, from the more to the less constrained one. This paper can also be considered as a survey on consistency criteria for DSMs. Key-words: shared memory, distributed system, consistency criteria, partial order. (R'esum'e : tsvp) IRISA, Campus de Beaulieu, 35042 Rennes-C'...

A general purpose parallel programmingmodel called mixed consistency is developed for distributed shared memory systems. This model combines two kinds of weak memory consistency conditions: causal memory and pipelined random access memory, and provides four kinds of explicit synchronization operations: read locks, write locks, barriers, and await operations. The resulting suite of memory and synchronization operations can be tailored to solve most programming problems in an efficient manner. Conditions are also developed under which the net effect of programming in this model is the same as programming with sequentially consistent memory. Several examples are included to illustrate the model and the correctness conditions. Keywords: distributed shared memory, memory consistency, concurrency, synchronization.

This paper presents and proves correct a distributed algorithm that implements a sequentially consistent collection of shared read/update objects. This algorithm is a generalization of one used in the Orca shared object system. The algorithm caches objects in the local memory of processors according to application needs; each read operation accesses a single copy of the object, while each update accesses all copies. The algorithm uses broadcast communication when it sends messages to replicated copies of an object, and it uses point-to-point communication when a message is sent to a single copy, and when a reply is returned. Copies of all the objects are kept consistent using a strategy based on sequence numbers for broadcasts. The algorithm is presented in two layers. The lower layer uses the given broadcast and point-to-point communication services, plus sequence numbers, to provide a new communication service called a context multicast channel. The higher layer uses a context multi...

Sequential consistency and causal consistency constitute two of the main consistency criteria used to define the semantics of accesses in the shared memory model. An execution is sequentially consistent if all processes can agree on a same legal sequential history of all the accesses; if processes perceive distinct legal sequential histories of all the accesses, the execution is only causally consistent (legality means that a read does not get an overwritten value). This paper studies synchronization constraints that, when obeyed by operations of a given causally consistent execution, make it sequentially consistent. More precisely, the paper introduces the MSC synchronization (mixed synchronization constraint) which generalizes (1) the known DRF (data race free) and CWF (concurrent write free) synchronizations and (2) a new one called CRF (concurrent read free). The MSC synchronization allows for concurrent conflicting operations on a same object, while ensuring sequential consiste...