Fault-Tolerant Average Execution Time Optimization for General Purpose Multi-Processor System-on Chips

Fault-tolerance is due to the semiconductor technology development important, not only for safety-critical systems but also for general-purpose (non-safety critical) systems. However, instead of guaranteeing that deadlines always are met, it is for general-purpose systems important to minimize the average execution time (AET) while ensuring fault-tolerance. For a given job and a soft (transient) error probability, we define mathematical formulas for AET that includes bus communication overhead for both voting (active replication) and rollback-recovery with checkpointing (RRC). And, for a given multi-processor system-on-chip (MPSoC), we define integer linear programming (ILP) models that minimize AET including bus communication overhead when: (1) selecting the number of checkpoints when using RRC, (2) finding the number of processors and job-to-processor assignment when using voting, and (3) defining fault-tolerance scheme (voting or RRC) per job and defining its usage for each job. Experiments demonstrate significant savings in AET.

An optimal weighted-average congestion based pricing scheme for enhanced QoS, , LNCS 4882, 2007

Pricing is an effective tool to control congestion and achieve quality of service (QoS) provisioning for multiple differentiated levels of service. In this paper, we consider the problem of pricing for congestion control in the case of a network of nodes under a single service class and multiple queues, and present a multi-layered pricing scheme. We propose an algorithm for finding the optimal state dependent price levels for individual queues, at each node. The pricing policy used depends on a weighted average queue length at each node. This helps in reducing frequent price variations and is in the spirit of the random early detection (RED) mechanism used in TCP/IP networks. We observe in our numerical results a considerable improvement in performance using our scheme over that of a recently proposed related scheme in terms of both throughput and delay performance. In particular, our approach exhibits a throughput improvement in the range of 34 to 69 percent in all cases studied (over all routes) over the above scheme.

Capacity Optimizing Hop Distance in a Mobile Ad Hoc Network with Power Control

In a dense multi-hop network of mobile nodes capable of applying adaptive power control, we consider the problem of finding the optimal hop distance that maximizes a certain throughput measure in bit-metres/sec, subject to average network power constraints. The mobility of nodes is restricted to a circular periphery area centered at the nominal location of nodes. We incorporate only randomly varying path-loss characteristics of channel gain due to the random motion of nodes, excluding any multi-path fading or shadowing effects. Computation of the throughput metric in such a scenario leads us to compute the probability density function of random distance between points in two circles. Using numerical analysis we discover that choosing the nearest node as next hop is not always optimal. Optimal throughput performance is also attained at non-trivial hop distances depending on the available average network power.

Reinforcement learning with average cost for adaptive control of traffic lights at intersections

We propose for the first time two reinforcement learning algorithms with function approximation for average cost adaptive control of traffic lights. One of these algorithms is a version of Q-learning with function approximation while the other is a policy gradient actor-critic algorithm that incorporates multi-timescale stochastic approximation. We show performance comparisons on various network settings of these algorithms with a range of fixed timing algorithms, as well as a Q-learning algorithm with full state representation that we also implement. We observe that whereas (as expected) on a two-junction corridor, the full state representation algorithm shows the best results, this algorithm is not implementable on larger road networks. The algorithm PG-AC-TLC that we propose is seen to show the best overall performance.

Average variation L2-stability criteria for time-varying feedback systems-A unified approach

The problem of developing L2-stability criteria for feedback systems with a single time-varying gain, which impose average variation constraints on the gain is treated. A unified approach is presented which facilitates the development of such average variation criteria for both linear and nonlinear systems. The stability criteria derived here are shown to be more general than the existing results.

New Insights into Optimal Discrete Rate Adaptation for Average Power Constrained Single and Multi-Node Systems

The throughput-optimal discrete-rate adaptation policy, when nodes are subject to constraints on the average power and bit error rate, is governed by a power control parameter, for which a closed-form characterization has remained an open problem. The parameter is essential in determining the rate adaptation thresholds and the transmit rate and power at any time, and ensuring adherence to the power constraint. We derive novel insightful bounds and approximations that characterize the power control parameter and the throughput in closed-form. The results are comprehensive as they apply to the general class of Nakagami-m (m >= 1) fading channels, which includes Rayleigh fading, uncoded and coded modulation, and single and multi-node systems with selection. The results are appealing as they are provably tight in the asymptotic large average power regime, and are designed and verified to be accurate even for smaller average powers.

Ultrasound modulation of coherent light in a multiple-scattering medium: experimental verification of nonzero average phase carried by light

We demonstrate the phase fluctuation introduced by oscillation of scattering centers in the focal volume of an ultrasound transducer in an optical tomography experiment has a nonzero mean. The conditions to be met for the above are: (i) the frequency of the ultrasound should be in the vicinity of the most dominant natural frequency of vibration of the ultrasound focal volume, (ii) the corresponding acoustic wavelength should be much larger than l(n)*, a modified transport mean-free-path applicable for phase decorrelation and (iii) the focal volume of the ultrasound transducer should not be larger than 4 - 5 times (l(n)*)(3). We demonstrate through simulations that as the ratio of the ultrasound focal volume to (l(n)*)(3) increases, the average of the phase fluctuation decreases and becomes zero when the focal volume becomes greater than around 4(l(n)*)(3); and through simulations and experiments that as the acoustic frequency increases from 100 Hz to 1 MHz, the average phase decreases to zero. Through experiments done in chicken breast we show that the average phase increases from around 110 degrees to 130 degrees when the background medium is changed from water to glycerol, indicating that the average of the phase fluctuation can be used to sense changes in refractive index deep within tissue.

A weighted average-based external clock synchronisation protocol for wireless sensor networks

Wireless Sensor Networks (WSNs) have many application scenarios where external clock synchronisation may be required because a WSN may consist of components which are not connected to each other. In this paper, we first propose a novel weighted average-based internal clock synchronisation (WICS) protocol, which synchronises all the clocks of a WSN with the clock of a reference node periodically. Based on this protocol, we then propose our weighted average-based external clock synchronisation (WECS) protocol. We have analysed the proposed protocols for maximum synchronisation error and shown that it is always upper bounded. Extensive simulation studies of the proposed protocols have been carried out using Castalia simulator. Simulation results validate our above theoretical claim and also show that the proposed protocols perform better in comparison to other protocols in terms of synchronisation accuracy. A prototype implementation of the WICS protocol using a few TelosB motes also validates the above conclusions.

Relay Selection for Geographical Forwarding in Sleep-Wake Cycling Wireless Sensor Networks

Our work is motivated by geographical forwarding of sporadic alarm packets to a base station in a wireless sensor network (WSN), where the nodes are sleep-wake cycling periodically and asynchronously. We seek to develop local forwarding algorithms that can be tuned so as to tradeoff the end-to-end delay against a total cost, such as the hop count or total energy. Our approach is to solve, at each forwarding node enroute to the sink, the local forwarding problem of minimizing one-hop waiting delay subject to a lower bound constraint on a suitable reward offered by the next-hop relay; the constraint serves to tune the tradeoff. The reward metric used for the local problem is based on the end-to-end total cost objective (for instance, when the total cost is hop count, we choose to use the progress toward sink made by a relay as the reward). The forwarding node, to begin with, is uncertain about the number of relays, their wake-up times, and the reward values, but knows the probability distributions of these quantities. At each relay wake-up instant, when a relay reveals its reward value, the forwarding node's problem is to forward the packet or to wait for further relays to wake-up. In terms of the operations research literature, our work can be considered as a variant of the asset selling problem. We formulate our local forwarding problem as a partially observable Markov decision process (POMDP) and obtain inner and outer bounds for the optimal policy. Motivated by the computational complexity involved in the policies derived out of these bounds, we formulate an alternate simplified model, the optimal policy for which is a simple threshold rule. We provide simulation results to compare the performance of the inner and outer bound policies against the simple policy, and also against the optimal policy when the source knows the exact number of relays. Observing the good performance and the ease of implementation of the simple policy, we apply it to our motivating problem, i.e., local geographical routing of sporadic alarm packets in a large WSN. We compare the end-to-end performance (i.e., average total delay and average total cost) obtained by the simple policy, when used for local geographical forwarding, against that obtained by the globally optimal forwarding algorithm proposed by Kim et al. 1].

Extended histories: improving regularity and performance in correlation prefetchers

Data Prefetchers identify and make use of any regularity present in the history/training stream to predict future references and prefetch them into the cache. The training information used is typically the primary misses seen at a particular cache level, which is a filtered version of the accesses seen by the cache. In this work we demonstrate that extending the training information to include secondary misses and hits along with primary misses helps improve the performance of prefetchers. In addition to empirical evaluation, we use the information theoretic metric entropy, to quantify the regularity present in extended histories. Entropy measurements indicate that extended histories are more regular than the default primary miss only training stream. Entropy measurements also help corroborate our empirical findings. With extended histories, further benefits can be achieved by triggering prefetches during secondary misses also. In this paper we explore the design space of extended prefetch histories and alternative prefetch trigger points for delta correlation prefetchers. We observe that different prefetch schemes benefit to a different extent with extended histories and alternative trigger points. Also the best performing design point varies on a per-benchmark basis. To meet these requirements, we propose a simple adaptive scheme that identifies the best performing design point for a benchmark-prefetcher combination at runtime. In SPEC2000 benchmarks, using all the L2 accesses as history for prefetcher improves the performance in terms of both IPC and misses reduced over techniques that use only primary misses as history. The adaptive scheme improves the performance of CZone prefetcher over Baseline by 4.6% on an average. These performance gains are accompanied by a moderate reduction in the memory traffic requirements.

A weighted average based external clock synchronization protocol for wireless sensor networks

Clock synchronization is an extremely important requirement of wireless sensor networks(WSNs). There are many application scenarios such as weather monitoring and forecasting etc. where external clock synchronization may be required because WSN itself may consists of components which are not connected to each other. A usual approach for external clock synchronization in WSNs is to synchronize the clock of a reference node with an external source such as UTC, and the remaining nodes synchronize with the reference node using an internal clock synchronization protocol. In order to provide highly accurate time, both the offset and the drift rate of each clock with respect to reference node are estimated from time to time, and these are used for getting correct time from local clock reading. A problem with this approach is that it is difficult to estimate the offset of a clock with respect to the reference node when drift rate of clocks varies over a period of time. In this paper, we first propose a novel internal clock synchronization protocol based on weighted averaging technique, which synchronizes all the clocks of a WSN to a reference node periodically. We call this protocol weighted average based internal clock synchronization(WICS) protocol. Based on this protocol, we then propose our weighted average based external clock synchronization(WECS) protocol. We have analyzed the proposed protocols for maximum synchronization error and shown that it is always upper bounded. Extensive simulation studies of the proposed protocols have been carried out using Castalia simulator. Simulation results validate our theoretical claim that the maximum synchronization error is always upper bounded and also show that the proposed protocols perform better in comparison to other protocols in terms of synchronization accuracy. A prototype implementation of the proposed internal clock synchronization protocol using a few TelosB motes also validates our claim.

Feedback overhead-aware fast distributed selection scheme for multi-node wireless systems

Opportunistic selection is a practically appealing technique that is used in multi-node wireless systems to maximize throughput, implement proportional fairness, etc. However, selection is challenging since the information about a node's channel gains is often available only locally at each node and not centrally. We propose a novel multiple access-based distributed selection scheme that generalizes the best features of the timer scheme, which requires minimal feedback but does not always guarantee successful selection, and the fast splitting scheme, which requires more feedback but guarantees successful selection. The proposed scheme's design explicitly accounts for feedback time overheads unlike the conventional splitting scheme and guarantees selection of the user with the highest metric unlike the timer scheme. We analyze and minimize the average time including feedback required by the scheme to select. With feedback overheads, the proposed scheme is scalable and considerably faster than several schemes proposed in the literature. Furthermore, the gains increase as the feedback overhead increases.

Tradeoff of average power and average delay for a point-to-point link with fading

We consider a discrete time system with packets arriving randomly at rate lambda per slot to a fading point-to-point link, for which the transmitter can control the number of packets served in a slot by varying the transmit power. We provide an asymptotic characterization of the minimum average delay of the packets, when average transmitter power is a small positive quantity V more than the minimum average power required for queue stability. We show that the minimum average delay will grow either as log (1/V) or 1/V when V down arrow 0, for certain sets of values of lambda. These sets are determined by the distribution of fading gain, the maximum number of packets which can be transmitted in a slot, and the assumed transmit power function, as a function of the fading gain and the number of packets transmitted. We identify a case where the above behaviour of the tradeoff differs from that obtained from a previously considered model, in which the random queue length process is assumed to evolve on the non-negative real line.

Tradeoff of average service cost and average delay for the state dependent M/M/1 queue

The optimal tradeoff between average service cost rate and average delay, is addressed for a M/M/1 queueing model with queue-length dependent service rates, chosen from a finite set. We provide an asymptotic characterization of the minimum average delay, when the average service cost rate is a small positive quantity V more than the minimum average service cost rate required for stability. We show that depending on the value of the arrival rate, the assumed service cost rate function, and the possible values of the service rates, the minimum average delay either a) increases only to a finite value, b) increases without bound as log(1/V), or c) increases without bound as 1/V, when V down arrow 0. We apply the analysis to a flow-level resource allocation model for a wireless downlink. We also investigate the asymptotic tradeoff for a sequence of policies which are obtained from an approximate fluid model for the M/M/1 queue.