We present an optimal real-time scheduling algorithm for multiprocessors $one that satisfies all task deadlines, when the total utilization demand does not exceed capacity of processors. The called LLREF, is designed based on a novel abstraction reasoning about execution behavior multiprocessors: time and local domain plane (or T-L plane). LLREF fluid model fairness notion, uses to describe schedules without using quanta, unlike Pfair (which quanta). show can be viewed as repeatedly...

This paper studies an extension of O'Hearn's incorrectness logic (IL) that allows backwards reasoning. IL in its current form does not generically permit We show this can be mitigated by extending with underspecification. The resulting combines underspecification (the result, or postcondition, only needs to formulate constraints over relevant variables) underapproximation (it focus on fewer than all the paths). prove soundness proof system, as well completeness for a defined subset...

We present heterogenous quorum-based asynchronous wake-up scheduling schemes for wireless sensor networks. The can ensure that two nodes adopt different quorum systems as their schedules hear each other at least once in bounded time intervals. propose such schemes: cyclic system pair (cqs-pair) and grid (gqs-pair). cqs-pair which contains provides an optimal solution, terms of energy saving ratio, scheduling. To quickly assemble a cqs-pair, we fast construction scheme is based on the...

A surveillance system, which tracks mobile targets, is one of the most important applications wireless sensor networks. When nodes operate in a duty cycling mode, tracking performance can be improved if target motion predicted and along trajectory proactively awakened. However, this will negatively influence energy efficiency constrain benefits cycling. In paper, we present Probability-based Prediction Sleep Scheduling protocol (PPSS) to improve proactive wake up. We start with designing...

The recent possibility of integrating multiple-OS-capable, high-core-count, heterogeneous-ISA processors in the same platform poses a question: given tight integration between system components, can shared memory programming model be adopted, enhancing programmability? If this done, an enormous amount existing code written for architectures would not have to rewritten use new paradigm (e.g., offloading) that is often very expensive and error prone. We propose software architecture composed...

We argue that the key underpinning of current state-of-the real-time practice - priority artifact and art deadline-based timeliness optimality are entirely inadequate for specifying objectives, reasoning about behavior, performing resource management can dependably satisfy objectives in many dynamic systems. time/utility functions utility accrual scheduling paradigm provide a more generalized, adaptive, flexible approach. Recent research has significantly advanced state-of-the-art paradigm....

Unikernels are minimal single-purpose virtual machines. They highly popular in the research domain due to benefits they provide. A barrier their widespread adoption is difficulty/impossibility port existing applications current unikernels. HermiTux first unikernel providing binary-compatibility with Linux applications. It composed of a hypervisor and lightweight kernel layer emulating OS interfaces at load- runtime accordance ABI. relieves application developers from burden porting software,...

Unikernels are minimal, single-purpose virtual machines. This new operating system model promises numerous benefits within many application domains in terms of lightweightness, performance, and security. Although the isolation between unikernels is generally recognized as strong, there no a unikernel itself. due to use single, unprotected address space, basic principle that provide their lightweightness performance benefits. In this paper, we propose design brings memory inside instance...

Edge computing is a recent paradigm that brings cloud services closer to the client. Among other features, edge offers extremely low client/server latencies. To consistently provide such latencies, need run on nodes are physically as close possible their clients. Thus, when client changes its physical location, service should migrate between maintain proximity. Differently from nodes, built with CPUs of different Instruction Set Architectures (ISAs), hence server program natively compiled...

We present a MAC-layer, soft real-time packet scheduling algorithm called UPA. UPA considers message model where packets have end-to-end timeliness requirements that are specified using Jensen's time-utility functions (TUFs). The seeks to maximize system-wide, aggregate utility. Since this problem is NP-hard, heuristically computes schedules with quadratic worst-case cost, faster than the previously best CMA algorithm. Our simulation studies show performs same as or significantly better for...

This paper presents a uni-processor real-time scheduling algorithm called the generic utility (which we refer to simply as GUS). GUS solves previously open problem-scheduling application activities that have time constraints specified using arbitrarily shaped time/utility functions and mutual exclusion resource constraints. A time/ function are constraint specification describes an activity's system of completion time. Given such constraints, consider objective maximizing total is accrued by...

We implement and experimentally evaluate the timeliness energy consumption behaviors of fourteen Real-Time Dynamic Voltage Frequency Scaling (RT-DVFS) schedulers on two hardware platforms. The include CC-EDF, LA-EDF, REUA, DRA, AGR1, among others, platforms Intel i5 processor AMD Zacate processor. Our studies reveal that measuring CPU power as cube frequency -- often done in simulation-based RT-DVFS literature ignores idle state consumption, which is significantly smaller than active...

Energy efficiency is one of the most important design considerations in running modern datacenters. Datacenter operating systems rely on software techniques such as execution migration to achieve energy across pools machines. Execution possible datacenters today because they consist mainly homogeneous-ISA However, recent market trends indicate that alternate ISAs ARM and PowerPC are pushing into datacenter, meaning current no longer applicable. How can be applied future heterogeneous-ISA datacenters?

New multi-leader consensus protocols leverage the Generalized Consensus specification to enable low latency, even load balancing, and high parallelism. However, these introduce inherent costs with significant performance impact: they need quorums bigger than minimum required solve track dependency relations among proposals. In this paper we present M2PAXOS, an implementation of that provides fast decisions (i.e., delivery a command in two communication delays) by leveraging composed majority...

Time constrained systems which operate in dynamic environments may have unknown worst case scenarios, large variances the sizes of data and event sets that they process (and thus, execution latencies resource requirements), not be statically characterizable, even by time invariant statistical distributions. The paper presents a specification language for describing environment dependent features. Also presented is an abstract model constructed (statically) from specifications, augmented...

We present a framework, called meta scheduler, for implementing real-time scheduling algorithms. The scheduler is portable middleware layer component designed implementations over POSIX-compliant operating systems. It accommodates pluggable algorithms while offering the flexibility of platform independence - singular underlying OS requirement now nearly ubiquitous POSIX compliance. versatility schedulers positions deployment in an interoperable heterogeneous environment. design and outline...

We revisit the shortest path problem in asynchronous duty-cycled wireless sensor networks, which exhibit time-dependent features. model time-varying link cost and distance from each node to sink as periodic functions. show that time-cost function satisfies FIFO property, makes solvable polynomial-time. Using ß-synchronizer, we propose a fast distributed algorithm build all-to-one paths with polynomial message complexity time complexity. The determines for all discrete times single execution,...

Distributed Transactional Memory (DTM) is a recent but promising model for programming distributed systems. It aims to present programmers with simple use concurrency control abstraction (transactions), while maintaining performance and scalability similar fine-grained locks. Any complications usually associated such locks (e.g., deadlocks) are avoided. We propose new DTM framework the Java Virtual Machine named Hyflow2. implement Hyflow2 in Scala base it on existing ScalaSTM API soon be...

This paper proposes CAESAR, a novel multi-leader Generalized Consensus protocol for geographically replicated sites. The main goal of CAESAR is to overcome one the major limitations existing approaches, which significant performance degradation when application workload produces conflicting requests. does that by changing way fast decision taken: its ordering not reject client request if quorum nodes reply with different dependency sets request. effectiveness demonstrated through an...

Herlihy and Koskinen's transactional boosting methodology addressed the challenge of converting concurrent data structures into ones. We present an optimistic for collections. Optimistic allows greater structure-specific optimizations, easier integration with STM frameworks, lower restrictions on boosted operations than original methodology.

Binary-level pointer analysis can be of use in symbolic execution, testing, verification, and decompilation software binaries. In various such contexts, it is crucial that the result trustworthy, i.e., formally established designations are overapproximative. This paper presents an approach to proven correct binary-level analysis. A salient property our first generically considers what proof obligations a generic abstract domain for must satisfy. allows easy instantiation different domains,...

We present HyFlow --- a distributed software transactional memory (D-STM) framework for concurrency control. is Java D-STM, with pluggable support directory lookup protocols, synchronization and recovery mechanisms, contention management policies, cache coherence network communication protocols. exports simple programming model that excludes locks: using (Java 5) annotations, atomic sections are defined as transactions, in which reads writes to shared, local remote objects appear take effect...