In order to analyze a program that involves pointers, it is necessary have (safe) information about what each pointer points to. There are many different approaches computing points-to information. This paper addresses techniques for flow- and context-insensitive interprocedural analysis of stack-based storage.The makes two contributions work in this area: - The first contribution set experiments explore the trade-offs between previously defined by Lars Andersen Bjarne Steensgaard. former...

A commutative replicated data type (CRDT) is one where all concurrent operations commute. The replicas of a CRDT converge automatically, without complex concurrency control. This paper describes Treedoc, novel design for cooperative text editing. An essential property that the identifiers Treedoc atoms are selected from dense space. We discuss practical alternatives implementing identifier space based on an extended binary tree. also storage and meta-data, mechanisms compacting In best case,...

Geo-replicated storage systems are at the core of current Internet services. The designers replication protocols used by these must choose between either supporting low-latency, eventually-consistent operations, or ensuring strong consistency to ease application correctness. We propose an alternative model, Explicit Consistency, that strengthens eventual with a guarantee preserve specific invariants defined applications. Given application-specific invariants, system supports Consistency...

Large-scale distributed systems often rely on replicated databases that allow a programmer to request different data consistency guarantees for operations, and thereby control their performance. Using such is far from trivial: requesting stronger in too many places may hurt performance, it few violate correctness. To help programmers this task, we propose the first proof rule establishing particular choice of various operations database enough ensure preservation given integrity invariant....

Developers of cloud-scale applications face a difficult decision which kind storage to use, summarised by the CAP theorem. Currently choice is between classical CP databases, provide strong guarantees but are slow, expensive, and unavailable under partition, NoSQL-style AP fast available, too hard program against. We present an alternative: Cure provides highest level that remains compatible with availability. These include: causal consistency (no ordering anomalies), atomicity (consistent...

We describe a novel approach to log-based reconciliation called IceCube. It is general and parameterised by application object semantics. IceCube considers more flexible orderings designed ease the burden of on programmers. captures static dynamic constraints between all pairs actions, proposes schedules that satisfy constraints, validates them against constraints.

Client-side apps (e.g., mobile or in-browser) need cloud data to be available in a local cache, for both reads and updates. For optimal user experience developer support, the cache should consistent fault-tolerant. In order scale high numbers of unreliable resource-poor clients, large database, system needs use resources sparingly. The SwiftCloud distributed object database is first provide fast writes via causally-consistent client-side backed by cloud. It thrifty scales well, thanks...

Modern cloud systems are geo-replicated to improve application latency and availability. Transactional consistency is essential for developers; however, the corresponding concurrency control commitment protocols costly in a setting. To minimize this cost, we identify following scalability properties: (i) only replicas updated by transaction T make steps execute T; (ii) read-only never waits concurrent transactions always commits; (iii) may read object versions committed after it started;...

We study a family of implementations for linked lists using fine-grain synchronisation. This approach enables greater concurrency, but correctness is challenge than classical, coarse-grain Our examples are demonstrative common design patterns such as lock coupling, optimistic, and lazy Although they highly concurrent, we prove that linearisable, safe, correctly implement high-level abstraction. proofs illustrate the power applicability rely-guarantee reasoning, well some its limitations. The...

Large-scale multicore architectures create new challenges for garbage collectors (GCs). In particular, throughput-oriented stop-the-world algorithms demonstrate good performance with a small number of cores, but have been shown to degrade badly beyond approximately 8 cores on 48-core OpenJDK 7. This negative result raises the question whether design has intrinsic limitations that would require radically different approach. Our study suggests answer is no, and there no compelling scalability...

On contemporary cache-coherent Non-Uniform Memory Access (ccNUMA) architectures, applications with a large memory footprint suffer from the cost of garbage collector (GC), because, as GC scans reference graph, it makes many remote accesses, saturating interconnect between nodes. We address this problem NumaGiC, mostly-distributed design. In order to maximise access locality during collection, thread avoids accessing different node, instead notifying message; nonetheless, NumaGiC drawbacks...

Navigational avoidance as a contrail mitigation strategy has the potential to reduce climate impact of aviation by much half. The effective implementation strategies requires forecasts state upper troposphere and lower stratosphere that are stable (consistent across range lead times) accurate (true reality). However, optimal criteria for evaluating whether forecasting system is sufficiently remain unclear. Here, we argue forecast stability best evaluated holistically asking estimated...

Client-side logic and storage are increasingly used in web mobile applications to improve response time availability. Current approaches tend be ad-hoc poorly integrated with the server-side logic. We present a principled approach integrate client-and storage. support both mergeable strongly consistent transactions that target either client or server replicas provide access causally-consistent snapshots efficiently. In presence of infrastructure faults, client-assisted failover solution...