We present the kernel language approach; a new way to teach programming that situates most of the... more We present the kernel language approach; a new way to teach programming that situates most of the widely known programming paradigms (including imperative, object-oriented, concurrent, logic, and functional) into a uniform setting that shows their deep relationships and how to use them together. Widely different practical languages (exemplified by Java, Haskell, Prolog, and Erlang) with their rich panoplies of abstractions and syntax are explained by straightforward translations into closely related kernel languages, simple languages that consist of small numbers of programmer-significant concepts. Kernel languages are easy to understand and have a simple formal semantics that can be used by practicing programmers to reason about correctness and complexity.
Electronic Proceedings in Theoretical Computer Science
Designing distributed systems to have predictable performance under high load is difficult becaus... more Designing distributed systems to have predictable performance under high load is difficult because of resource exhaustion, non-linearity, and stochastic behaviour. Timeliness, i.e., delivering results within defined time bounds, is a central aspect of predictable performance. In this paper, we focus on timeliness using the ∆Q Systems Development paradigm (∆QSD, developed by PNSol), which computes timeliness by modelling systems observationally using so-called outcome expressions. An outcome expression is a compositional definition of a system's observed behaviour in terms of its basic operations. Given the behaviour of the basic operations, ∆QSD efficiently computes the stochastic behaviour of the whole system including its timeliness. This paper formally proves useful algebraic properties of outcome expressions w.r.t. timeliness. We prove the different algebraic structures the set of outcome expressions form with the different ∆QSD operators and demonstrate why those operators do not form richer structures. We prove or disprove the set of all possible distributivity results on outcome expressions. On our way for disproving 8 of those distributivity results, we develop a technique called properisation, which gives rise to the first body of maths for improper random variables. Finally, we also prove 14 equivalences that have been used in the past in the practice of ∆QSD. An immediate benefit is rewrite rules that can be used for design exploration under established timeliness equivalence. This work is part of an ongoing project to disseminate and build tool support for ∆QSD. The ability to rewrite outcome expressions is essential for efficient tool support. Algebraic Reasoning About Timeliness This paper defines and proves algebraic properties of the ∆QSD operators w.r.t. timeliness, i.e., delivering outcomes within the acceptable time-frames. In this paper, our sole resource of concern is time, although ∆QSD includes other types of resources and their interaction. This theoretical work is part of an ongoing project to disseminate and build tool support for ∆QSD, to make it available to the wide community of system engineers. We base our work on the ∆QSD formalisation given by Haeri et al. [11], which defines outcome expressions and their semantics, and gives a real-world example of ∆QSD taken from the blockchain domain. Contributions This paper gives a firm mathematical foundation for ∆QSD, and uses this to establish important algebraic properties of the ∆QSD operators with respect to timeliness, i.e., when the relevant resource is time. This paper is based on a general model theory of resource analysis for systems specified using outcome expressions [12]. That model theory is the first of its kind and we specialise it using the timeliness analysis recipe that is commonly used in ∆QSD (Definition 3). • We show that the set of outcome expressions forms different algebraic structures with the different ∆QSD operators (Theorems 1-4). • We establish 3 distributivity results in Section 7 about the ∆QSD operators (Theorem 6). • We rule out the formation of certain richer algebraic structures by the set of outcome expressions and the current ∆QSD operators (Remarks 2, 3, and 4).
We present the current state of a new implementation of the constraint engine for the Mozart prog... more We present the current state of a new implementation of the constraint engine for the Mozart programming system. Our implementation integrates the Gecode constraint library into the core of Mozart version 2.0. Doing so, we allow users to take advantage of the efficiency of Gecode propagators transparently by maintaining the existing language constraints abstractions. Future Mozart systems can thus benefit from the rapid pace of constraint solving optimizations that are included in each new Gecode version. We use two well-known puzzles to illustrate the system and present all available abstractions.
2017 IEEE International Conference on Edge Computing (EDGE), 2017
Distributed applications will break down or perform poorly when there are too many failures (of n... more Distributed applications will break down or perform poorly when there are too many failures (of nodes and/or communication) in the operating environment. Failures happen frequently on edge networks including mobile and ad hoc networks, but are also unexpectedly common on the general Internet. We propose an approach for designing stressaware distributed applications that can take environment stress into account to improve their behavior. We give a concrete illustration of the approach by targeting applications built on top of a Structured Overlay Network (SON). Our underlying SON is Reversible and Phase-Aware. A system is Reversible if the set of operations it provides is a function (called the reversibility function) of its current stress (i.e., all perturbing effects of the environment, including faults), and does not depend on past stress. Reversibility generalizes standard fault tolerance with nested fault models. When the fault rate goes outside the scope of one model, then it is still inside the next one. In order to approximate the reversibility function we introduce the concept of Phase, which is a per-node property that gives a qualitative measure of the available system operations under the current stress. Phase can be determined with no extra distributed operations. We show that making the phase available to applications allows them to improve their behavior in environments with high and variable stress. We propose a Phase API and we design an application, a collaborative graphic editor, that takes advantage of phase to enhance self-adaptation and selfoptimization properties. Furthermore, we analyze how the application itself can achieve reversibility in the applicationlevel semantics. Using the phase of the underlying node, the application provides an indication to the user regarding its behavior. Thus, the application has improved behavior with respect to the user, i.e., the user can better understand and decide what to do in a high-stress environment.
Much progress has been made in distributed computing in the areas of distribution structure, open... more Much progress has been made in distributed computing in the areas of distribution structure, open computing, fault tolerance, and security. Yet, writing distributed applications remains difficult because the programmer has to manage models of these areas explicitly. A major challenge is to integrate the four models into a coherent development platform. Such a platform should make it possible to cleanly separate an application's functionality from the other four concerns. Concurrent constraint programming, an evolution of concurrent logic programming, has both the expressiveness and the formal foundation needed to attempt this integration. As a first step, we have designed and built a platform that separates an application's functionality from its distribution structure. We have prototyped several collaborative tools with this platform, including a shared graphic editor whose design is presented in detail. The platform efficiently implements Distributed Oz, which extends the Oz language with constructs to express the distribution structure and with basic primitives for open computing, failure detection and handling, and resource control. Oz appears to the programmer as a concurrent object-oriented language with dataflow synchronization. Oz is based on a higher-order, state-aware, concurrent constraint computation model.
Abstract. As Internet applications become larger and more complex, the task of managing them beco... more Abstract. As Internet applications become larger and more complex, the task of managing them becomes overwhelming. Abnormal events such as software updates, failures, attacks, and hotspots become frequent. The SELFMAN project is tackling this problem by combining two tech- ...
Detachable user interfaces consist of graphical user interfaces whose parts or whole can be detac... more Detachable user interfaces consist of graphical user interfaces whose parts or whole can be detached at run-time from their host, migrated onto another computing platform while carrying out the task, possibly adapted to the new platform and attached to the target platform in a peer-to-peer fashion. Detaching is the property of splitting a part of a UI for transferring it onto another platform. Attaching is the reciprocal property: a part of an existing interface can be attached to the currently being used interface so as to recompose another one on-demand, according to user's needs, task requirements. Assembling interface parts by detaching and attaching allows dynamically composing, decomposing and re-composing new interfaces on demand. To support this interaction paradigm, a development infrastructure has been developed based on a series of primitives such as display, undisplay, copy, expose, return, transfer, delegate, and switch. We exemplify it with QTkDraw, a painting application with attaching and detaching based on the development infrastructure.Anglai
Proceedings of the ACM on Programming Languages, 2020
Oz is a programming language designed to support multiple programming paradigms in a clean factor... more Oz is a programming language designed to support multiple programming paradigms in a clean factored way that is easy to program despite its broad coverage. It started in 1991 as a collaborative effort by the DFKI (Germany) and SICS (Sweden) and led to an influential system, Mozart, that was released in 1999 and widely used in the 2000s for practical applications and education. We give the history of Oz as it developed from its origins in logic programming, starting with Prolog, followed by concurrent logic programming and constraint logic programming, and leading to its two direct precursors, the concurrent constraint model and the Andorra Kernel Language (AKL). We give the lessons learned from the Oz effort including successes and failures and we explain the principles underlying the Oz design. Oz is defined through a kernel language, which is a formal model similar to a foundational calculus, but that is designed to be directly useful to the programmer. The kernel language is orga...
We propose Lasp, a new programming model designed to simplify large-scale distributed programming... more We propose Lasp, a new programming model designed to simplify large-scale distributed programming. Lasp combines ideas from deterministic dataflow programming together with conflictfree replicated data types (CRDTs). This provides support for computations where not all participants are online together at a given moment. The initial design presented here provides powerful primitives for composing CRDTs, which lets us write long-lived faulttolerant distributed applications with nonmonotonic behavior in a monotonic framework. Given reasonable models of node-to-node communications and node failures, we prove formally that a Lasp program can be considered as a functional program that supports functional reasoning and programming techniques. We have implemented Lasp as an Erlang library built on top of the Riak Core distributed systems framework. We have developed one nontrivial large-scale application, the advertisement counter scenario from the SyncFree research project. We plan to extend our current prototype into a general-purpose language in which synchronization is used as little as possible.
There is no doubt about the increase in popularity of decentralised systems over the classical cl... more There is no doubt about the increase in popularity of decentralised systems over the classical client-server architecture in distributed applications. These systems are developed mainly as peer-to-peer networks where it is possible to observe many strategies to organise the peers. The most popular one for structured networks is the ring topology. Despite many advantages offered by this topology, the maintenance of the ring is very costly, being difficult to guarantee lookup consistency and fault-tolerance all the time. By increasing self-management in the system we are able to deal with these issues. We model ring maintenance as a selforganising and self-healing system using feedback loops. As a result, we introduce a novel relaxed-ring topology that is able to provide fault-tolerance with realistic assumptions concerning failure detection. Limitations related to failure handling are clearly identified, providing strong guarantees to develop applications on top of the relaxed-ring architecture. Besides permanent failures, the paper analyses temporary failures and broken links, which are often ignored.
We present a new programming model for largescale mobile and "Internet of Things" style distribut... more We present a new programming model for largescale mobile and "Internet of Things" style distributed applications. The model consists of two layers: a language layer based on the Lasp language with a runtime layer based on epidemic broadcast. The Lasp layer provides deterministic coordinationfree computation primitives based on conflict-free replicated data types (CRDTs). The epidemic broadcast layer is based on the Plumtree protocol. It provides a communication framework where clients may only have a partial view of membership and may not want to participate in or have knowledge of all active computations. We motivate the new model with a nontrivial mobile application, a distributed ad counter, and we give the model's formal semantics.
We propose a new "generic" abstraction for Erlang/OTP that aids in the implementation of dataflow... more We propose a new "generic" abstraction for Erlang/OTP that aids in the implementation of dataflow programming languages and models on the Erlang VM. This abstraction simplifies the implementation of "processing elements" in dataflow languages by providing a simple callback interface in the style of the gen server and gen fsm abstractions. We motivate the use of this new abstraction by examining the implementation of a distributed dataflow programming variant called Lasp.
Information systems → Distributed database transactions; Storage replication; Online services are... more Information systems → Distributed database transactions; Storage replication; Online services are often deployed over geographically-scattered data centers (geo-replication), which allows services to be highly available and reduces access latency. On the down side, to provide ACID transactions, global certification (i.e., across data centers) is needed to detect conflicts between concurrent transactions executing at different data centers. The global certification phase reduces throughput because transactions need to hold pre-commit locks, and it increases client-perceived latency because global certification lies in the critical path of transaction execution.
† This paper is an extended version of our paper published in 33 rd Symposium on Implementation a... more † This paper is an extended version of our paper published in 33 rd Symposium on Implementation and Application of Functional Languages (IFL21), as a draft.
This work presents Speculative Transaction Replication (STR), a protocol that exploits transparen... more This work presents Speculative Transaction Replication (STR), a protocol that exploits transparent speculation techniques to enhance performance of geo-distributed, partially replicated transactional data stores. In addition, we define a new consistency model, Speculative Snapshot Isolation (SPSI), that extends the semantics of Snapshot Isolation (SI) to shelter applications from the subtle anomalies that can arise from using speculative transaction processing. SPSI extends SI in an intuitive and rigorous fashion by specifying desirable atomicity and isolation guarantees that must hold when using speculative execution. STR provides a form of speculation that is fully transparent for programmers (it does not expose the effects of misspeculations to clients). Since the speculation techniques employed by STR satisfy SPSI, they can be leveraged by application programs in a transparent way, without requiring any source-code modification to applications designed to operate using SI. STR combines two key techniques: speculative reads, which allow transactions to observe pre-committed versions, which can reduce the 'effective duration' of pre-commit locks and enhance throughput; Precise Clocks, a novel timestamping mechanism that uses per-item timestamps with physical clocks, which together greatly enhance the probability of successful speculation. We assess STR's performance on up to nine geo-distributed Amazon EC2 data centers, using both synthetic benchmarks as well as realistic benchmarks (TPC-C and RUBiS). Our evaluation shows that STR achieves throughput gains up to 11× and latency reduction up to 10×, in workloads characterized by low inter-data center contention. Furthermore, thanks to a self-tuning mechanism that dynamically and transparently enables and disables speculation, STR offers robust performance even when faced with unfavourable workloads that suffer from high misspeculation rates.
We present the kernel language approach; a new way to teach programming that situates most of the... more We present the kernel language approach; a new way to teach programming that situates most of the widely known programming paradigms (including imperative, object-oriented, concurrent, logic, and functional) into a uniform setting that shows their deep relationships and how to use them together. Widely different practical languages (exemplified by Java, Haskell, Prolog, and Erlang) with their rich panoplies of abstractions and syntax are explained by straightforward translations into closely related kernel languages, simple languages that consist of small numbers of programmer-significant concepts. Kernel languages are easy to understand and have a simple formal semantics that can be used by practicing programmers to reason about correctness and complexity.
Electronic Proceedings in Theoretical Computer Science
Designing distributed systems to have predictable performance under high load is difficult becaus... more Designing distributed systems to have predictable performance under high load is difficult because of resource exhaustion, non-linearity, and stochastic behaviour. Timeliness, i.e., delivering results within defined time bounds, is a central aspect of predictable performance. In this paper, we focus on timeliness using the ∆Q Systems Development paradigm (∆QSD, developed by PNSol), which computes timeliness by modelling systems observationally using so-called outcome expressions. An outcome expression is a compositional definition of a system's observed behaviour in terms of its basic operations. Given the behaviour of the basic operations, ∆QSD efficiently computes the stochastic behaviour of the whole system including its timeliness. This paper formally proves useful algebraic properties of outcome expressions w.r.t. timeliness. We prove the different algebraic structures the set of outcome expressions form with the different ∆QSD operators and demonstrate why those operators do not form richer structures. We prove or disprove the set of all possible distributivity results on outcome expressions. On our way for disproving 8 of those distributivity results, we develop a technique called properisation, which gives rise to the first body of maths for improper random variables. Finally, we also prove 14 equivalences that have been used in the past in the practice of ∆QSD. An immediate benefit is rewrite rules that can be used for design exploration under established timeliness equivalence. This work is part of an ongoing project to disseminate and build tool support for ∆QSD. The ability to rewrite outcome expressions is essential for efficient tool support. Algebraic Reasoning About Timeliness This paper defines and proves algebraic properties of the ∆QSD operators w.r.t. timeliness, i.e., delivering outcomes within the acceptable time-frames. In this paper, our sole resource of concern is time, although ∆QSD includes other types of resources and their interaction. This theoretical work is part of an ongoing project to disseminate and build tool support for ∆QSD, to make it available to the wide community of system engineers. We base our work on the ∆QSD formalisation given by Haeri et al. [11], which defines outcome expressions and their semantics, and gives a real-world example of ∆QSD taken from the blockchain domain. Contributions This paper gives a firm mathematical foundation for ∆QSD, and uses this to establish important algebraic properties of the ∆QSD operators with respect to timeliness, i.e., when the relevant resource is time. This paper is based on a general model theory of resource analysis for systems specified using outcome expressions [12]. That model theory is the first of its kind and we specialise it using the timeliness analysis recipe that is commonly used in ∆QSD (Definition 3). • We show that the set of outcome expressions forms different algebraic structures with the different ∆QSD operators (Theorems 1-4). • We establish 3 distributivity results in Section 7 about the ∆QSD operators (Theorem 6). • We rule out the formation of certain richer algebraic structures by the set of outcome expressions and the current ∆QSD operators (Remarks 2, 3, and 4).
We present the current state of a new implementation of the constraint engine for the Mozart prog... more We present the current state of a new implementation of the constraint engine for the Mozart programming system. Our implementation integrates the Gecode constraint library into the core of Mozart version 2.0. Doing so, we allow users to take advantage of the efficiency of Gecode propagators transparently by maintaining the existing language constraints abstractions. Future Mozart systems can thus benefit from the rapid pace of constraint solving optimizations that are included in each new Gecode version. We use two well-known puzzles to illustrate the system and present all available abstractions.
2017 IEEE International Conference on Edge Computing (EDGE), 2017
Distributed applications will break down or perform poorly when there are too many failures (of n... more Distributed applications will break down or perform poorly when there are too many failures (of nodes and/or communication) in the operating environment. Failures happen frequently on edge networks including mobile and ad hoc networks, but are also unexpectedly common on the general Internet. We propose an approach for designing stressaware distributed applications that can take environment stress into account to improve their behavior. We give a concrete illustration of the approach by targeting applications built on top of a Structured Overlay Network (SON). Our underlying SON is Reversible and Phase-Aware. A system is Reversible if the set of operations it provides is a function (called the reversibility function) of its current stress (i.e., all perturbing effects of the environment, including faults), and does not depend on past stress. Reversibility generalizes standard fault tolerance with nested fault models. When the fault rate goes outside the scope of one model, then it is still inside the next one. In order to approximate the reversibility function we introduce the concept of Phase, which is a per-node property that gives a qualitative measure of the available system operations under the current stress. Phase can be determined with no extra distributed operations. We show that making the phase available to applications allows them to improve their behavior in environments with high and variable stress. We propose a Phase API and we design an application, a collaborative graphic editor, that takes advantage of phase to enhance self-adaptation and selfoptimization properties. Furthermore, we analyze how the application itself can achieve reversibility in the applicationlevel semantics. Using the phase of the underlying node, the application provides an indication to the user regarding its behavior. Thus, the application has improved behavior with respect to the user, i.e., the user can better understand and decide what to do in a high-stress environment.
Much progress has been made in distributed computing in the areas of distribution structure, open... more Much progress has been made in distributed computing in the areas of distribution structure, open computing, fault tolerance, and security. Yet, writing distributed applications remains difficult because the programmer has to manage models of these areas explicitly. A major challenge is to integrate the four models into a coherent development platform. Such a platform should make it possible to cleanly separate an application's functionality from the other four concerns. Concurrent constraint programming, an evolution of concurrent logic programming, has both the expressiveness and the formal foundation needed to attempt this integration. As a first step, we have designed and built a platform that separates an application's functionality from its distribution structure. We have prototyped several collaborative tools with this platform, including a shared graphic editor whose design is presented in detail. The platform efficiently implements Distributed Oz, which extends the Oz language with constructs to express the distribution structure and with basic primitives for open computing, failure detection and handling, and resource control. Oz appears to the programmer as a concurrent object-oriented language with dataflow synchronization. Oz is based on a higher-order, state-aware, concurrent constraint computation model.
Abstract. As Internet applications become larger and more complex, the task of managing them beco... more Abstract. As Internet applications become larger and more complex, the task of managing them becomes overwhelming. Abnormal events such as software updates, failures, attacks, and hotspots become frequent. The SELFMAN project is tackling this problem by combining two tech- ...
Detachable user interfaces consist of graphical user interfaces whose parts or whole can be detac... more Detachable user interfaces consist of graphical user interfaces whose parts or whole can be detached at run-time from their host, migrated onto another computing platform while carrying out the task, possibly adapted to the new platform and attached to the target platform in a peer-to-peer fashion. Detaching is the property of splitting a part of a UI for transferring it onto another platform. Attaching is the reciprocal property: a part of an existing interface can be attached to the currently being used interface so as to recompose another one on-demand, according to user's needs, task requirements. Assembling interface parts by detaching and attaching allows dynamically composing, decomposing and re-composing new interfaces on demand. To support this interaction paradigm, a development infrastructure has been developed based on a series of primitives such as display, undisplay, copy, expose, return, transfer, delegate, and switch. We exemplify it with QTkDraw, a painting application with attaching and detaching based on the development infrastructure.Anglai
Proceedings of the ACM on Programming Languages, 2020
Oz is a programming language designed to support multiple programming paradigms in a clean factor... more Oz is a programming language designed to support multiple programming paradigms in a clean factored way that is easy to program despite its broad coverage. It started in 1991 as a collaborative effort by the DFKI (Germany) and SICS (Sweden) and led to an influential system, Mozart, that was released in 1999 and widely used in the 2000s for practical applications and education. We give the history of Oz as it developed from its origins in logic programming, starting with Prolog, followed by concurrent logic programming and constraint logic programming, and leading to its two direct precursors, the concurrent constraint model and the Andorra Kernel Language (AKL). We give the lessons learned from the Oz effort including successes and failures and we explain the principles underlying the Oz design. Oz is defined through a kernel language, which is a formal model similar to a foundational calculus, but that is designed to be directly useful to the programmer. The kernel language is orga...
We propose Lasp, a new programming model designed to simplify large-scale distributed programming... more We propose Lasp, a new programming model designed to simplify large-scale distributed programming. Lasp combines ideas from deterministic dataflow programming together with conflictfree replicated data types (CRDTs). This provides support for computations where not all participants are online together at a given moment. The initial design presented here provides powerful primitives for composing CRDTs, which lets us write long-lived faulttolerant distributed applications with nonmonotonic behavior in a monotonic framework. Given reasonable models of node-to-node communications and node failures, we prove formally that a Lasp program can be considered as a functional program that supports functional reasoning and programming techniques. We have implemented Lasp as an Erlang library built on top of the Riak Core distributed systems framework. We have developed one nontrivial large-scale application, the advertisement counter scenario from the SyncFree research project. We plan to extend our current prototype into a general-purpose language in which synchronization is used as little as possible.
There is no doubt about the increase in popularity of decentralised systems over the classical cl... more There is no doubt about the increase in popularity of decentralised systems over the classical client-server architecture in distributed applications. These systems are developed mainly as peer-to-peer networks where it is possible to observe many strategies to organise the peers. The most popular one for structured networks is the ring topology. Despite many advantages offered by this topology, the maintenance of the ring is very costly, being difficult to guarantee lookup consistency and fault-tolerance all the time. By increasing self-management in the system we are able to deal with these issues. We model ring maintenance as a selforganising and self-healing system using feedback loops. As a result, we introduce a novel relaxed-ring topology that is able to provide fault-tolerance with realistic assumptions concerning failure detection. Limitations related to failure handling are clearly identified, providing strong guarantees to develop applications on top of the relaxed-ring architecture. Besides permanent failures, the paper analyses temporary failures and broken links, which are often ignored.
We present a new programming model for largescale mobile and "Internet of Things" style distribut... more We present a new programming model for largescale mobile and "Internet of Things" style distributed applications. The model consists of two layers: a language layer based on the Lasp language with a runtime layer based on epidemic broadcast. The Lasp layer provides deterministic coordinationfree computation primitives based on conflict-free replicated data types (CRDTs). The epidemic broadcast layer is based on the Plumtree protocol. It provides a communication framework where clients may only have a partial view of membership and may not want to participate in or have knowledge of all active computations. We motivate the new model with a nontrivial mobile application, a distributed ad counter, and we give the model's formal semantics.
We propose a new "generic" abstraction for Erlang/OTP that aids in the implementation of dataflow... more We propose a new "generic" abstraction for Erlang/OTP that aids in the implementation of dataflow programming languages and models on the Erlang VM. This abstraction simplifies the implementation of "processing elements" in dataflow languages by providing a simple callback interface in the style of the gen server and gen fsm abstractions. We motivate the use of this new abstraction by examining the implementation of a distributed dataflow programming variant called Lasp.
Information systems → Distributed database transactions; Storage replication; Online services are... more Information systems → Distributed database transactions; Storage replication; Online services are often deployed over geographically-scattered data centers (geo-replication), which allows services to be highly available and reduces access latency. On the down side, to provide ACID transactions, global certification (i.e., across data centers) is needed to detect conflicts between concurrent transactions executing at different data centers. The global certification phase reduces throughput because transactions need to hold pre-commit locks, and it increases client-perceived latency because global certification lies in the critical path of transaction execution.
† This paper is an extended version of our paper published in 33 rd Symposium on Implementation a... more † This paper is an extended version of our paper published in 33 rd Symposium on Implementation and Application of Functional Languages (IFL21), as a draft.
This work presents Speculative Transaction Replication (STR), a protocol that exploits transparen... more This work presents Speculative Transaction Replication (STR), a protocol that exploits transparent speculation techniques to enhance performance of geo-distributed, partially replicated transactional data stores. In addition, we define a new consistency model, Speculative Snapshot Isolation (SPSI), that extends the semantics of Snapshot Isolation (SI) to shelter applications from the subtle anomalies that can arise from using speculative transaction processing. SPSI extends SI in an intuitive and rigorous fashion by specifying desirable atomicity and isolation guarantees that must hold when using speculative execution. STR provides a form of speculation that is fully transparent for programmers (it does not expose the effects of misspeculations to clients). Since the speculation techniques employed by STR satisfy SPSI, they can be leveraged by application programs in a transparent way, without requiring any source-code modification to applications designed to operate using SI. STR combines two key techniques: speculative reads, which allow transactions to observe pre-committed versions, which can reduce the 'effective duration' of pre-commit locks and enhance throughput; Precise Clocks, a novel timestamping mechanism that uses per-item timestamps with physical clocks, which together greatly enhance the probability of successful speculation. We assess STR's performance on up to nine geo-distributed Amazon EC2 data centers, using both synthetic benchmarks as well as realistic benchmarks (TPC-C and RUBiS). Our evaluation shows that STR achieves throughput gains up to 11× and latency reduction up to 10×, in workloads characterized by low inter-data center contention. Furthermore, thanks to a self-tuning mechanism that dynamically and transparently enables and disables speculation, STR offers robust performance even when faced with unfavourable workloads that suffer from high misspeculation rates.
Uploads
Papers by Peter Van Roy