Saurabh Kumar Raina

Assessing the Impact of Partial Verifications against Silent Data Corruptions

Silent errors, or silent data corruptions, constitute a major threat on very large scale platforms. When a silent error strikes, it is not detected immediately but only after some delay, which prevents the use of pure periodic check pointing approaches devised for fail-stop errors. Instead, check pointing must be coupled with some verification mechanism to guarantee that corrupted data will never be written into the checkpoint file. Such a guaranteed verification mechanism typically incurs a high cost. In this paper, we assess the impact of using partial verification mechanisms in addition to a guaranteed verification. The main objective is to investigate to which extent it is worthwhile to use some light cost but less accurate verifications in the middle of a periodic computing pattern, which ends with a guaranteed verification right before each checkpoint. Introducing partial verifications dramatically complicates the analysis, but we are able to analytically determine the optimal computing pattern (up to the first-order approximation), including the optimal length of the pattern, the optimal number of partial verifications, as well as their optimal positions inside the pattern. Performance evaluations based on a wide range of parameters confirm the benefit of using partial verifications under certain scenarios, when compared to the baseline algorithm that uses only guaranteed verifications.

Efficient checkpoint/verification patterns

Errors have become a critical problem for high-performance computing. Checkpointing protocols are often used for error recovery after fail-stop failures. However, silent errors cannot be ignored, and their peculiarity is that such errors are identified only when the corrupted data is activated. To cope with silent errors, we need a verification mechanism to check whether the application state is correct. Checkpoints should be supplemented with verifications to detect silent errors. When a verification is successful, only the last checkpoint needs to be kept in memory because it is known to be correct. In this paper, we analytically determine the best balance of verifications and checkpoints so as to optimize platform throughput. We introduce a balanced algorithm using a pattern with p checkpoints and q verifications, which regularly interleaves both checkpoints and verifications across same-size computational chunks. We show how to compute the waste of an arbitrary pattern, and we prove that the balanced algorithm is optimal when the platform MTBF (mean time between failures) is large in front of the other parameters (checkpointing, verification and recovery costs). We conduct several simulations to show the gain achieved by this balanced algorithm for well-chosen values of p and q, compared with the base algorithm that always perform a verification just before taking a checkpoint (p = q = 1), and we exhibit gains of up to 19%.

Which Verification for Soft Error Detection

Many methods are available to detect silent errors in high-performance computing (HPC) applications. Each comes with a given cost and recall (fraction of all errors that are actually detected). The main contribution of this paper is to characterize the optimal computational pattern for an application: which detector(s) to use, how many detectors of each type to use, together with the length of the work segment that precedes each of them. We conduct a comprehensive complexity analysis of this optimization problem, showing NP-completeness and designing an FPTAS (Fully Polynomial-Time Approximation Scheme). On the practical side, we provide a greedy algorithm whose performance is shown to be close to the optimal for a realistic set of evaluation scenarios.

Distance and direction-based location aided multi-hop routing protocol for vehicular ad-hoc networks

Vehicular ad-hoc network VANET is an important technology for future development of intelligent transportation systems. Due to the highly-dynamic nature of vehicular nodes, network topology changes very frequently which complicate the routing of data packets. A number of routing protocols have been developed by various researchers. Hop count is a key parameter in evaluating the performance of a routing protocol. Number of hops required in directional routing protocols is more in comparison to other routing protocols. For a city vehicular traffic scenario we propose an improved direction-based location aided routing D-LAR protocol that we call distance and direction-based location aided routing DD-LAR protocol. A mathematical model for DD-LAR protocol has been developed to examine its performance on path duration and hop count metric. Simulations have been done in MATLAB and from the results, it can be seen that DD-LAR routing protocol shows significant performance improvement over D-LAR and LAR in terms of reducing the hop count and maximising the path duration.

Hop count analysis of location aided multihop routing protocols for VANETs

Vehicular ad-hoc network (VANET) is an important technology for future development of intelligent transportation systems. VANETs are being used as a tool for improving road safety by warning the drivers about accidents occurred ahead of them or for providing internet access to the passengers via gateways along the road. Although VANET is an application of MANETs but routing of data packets is more challenging in VANETs. This is because of frequent change in network topology due to the highly-dynamic nature of vehicular nodes. Researchers have developed various routing protocols for VANETs and hop count among others is considered as an important metric in evaluating these protocols. LAR (Location Aided Routing) and D-LAR (Directional-LAR) protocols are well known position based routing protocols for VANETs. In this paper we have proposed another protocol which is an important extension of D-LAR and which improves upon hop count in comparison to both LAR and D-LAR protocols. This protocol uses distance in addition to the directional routing strategy to select the next hop node, which results in a stable and shorter route. We have devised a generic HopCount Algorithm to compare the three protocols. Simulations are done to corroborate algorithmic analysis for multiple scenarios which depict dense environment. Performance results clearly depict that DD-LAR outperforms the other two protocols and with an increase in transmission range or vehicular density, hop count value is further reduced.

Performance analysis of routing protocols for vehicular adhoc networks using NS2/SUMO

Development of intelligent transportation system is the need of all the developing countries where urbanization and industrialization is rapidly growing. VANETs are being used as a tool for improving road safety by alarming the drivers about accidents occurred ahead of them or for providing internet access to the passengers via gateways along the road. Due to highly dynamic nature of nodes in VANETs, designing a routing protocol for VANET is quite challenging compared to MANET environment. Researchers have suggested several routing mechanism for VANETs. Few routing decision are based on topology based selection whereas others have considered different parameters like location information of nodes, traffic lights etc. As no benchmarking scheme is available for choosing a routing protocol in VANET, this article gives an insight on how to choose a routing protocol depends on varying condition of traffic. Three popular protocols AODV, DSR and LAR have been chosen for analysis on varying traffic environment. All the three protocols have been critically tested for different metrics such as Throughput, Packet Delivery Ratio and Routing overhead during the simulation. Simulation is carried out with the help of open-source simulation tools NS2, a network simulator, and SUMO, a traffic simulator.

SDLAR: Self Adaptive Density-Based Location-Aware Routing in Mobile Adhoc Network

Nowadays, location aware routing protocols are accepted potentially, scalable and efficient solution for routing in MANET. The advantages of these routing protocols are that they perform route discovery in a smaller region known as request zone, instead of doing route discovery in the entire network. This shows that the size and shape of the request zone play a major role to enhance the performance of routing procedure. Hence, the paper proposes an efficient scheme which focuses on creation and adjustment of size of request zone to find a stable path with less communication overhead. The protocol confines route discovery within an ellipse shaped request zone to reduce the routing overhead and achieve path stability. Further, the proposed protocol uses a density metric to resize the request zone for successful route discovery. Simulation results show that the proposed protocol can help to improve the path stability with lesser routing overhead than LAR-1.

Coping with recall and precision of soft error detectors

Abstract Many methods are available to detect silent errors in high-performance computing (HPC) applications. Each method comes with a cost, a recall (fraction of all errors that are actually detected, i.e., false negatives), and a precision (fraction of true errors amongst all detected errors, i.e., false positives). The main contribution of this paper is to characterize the optimal computing pattern for an application: which detector(s) to use, how many detectors of each type to use, together with the length of the work segment that precedes each of them. We first prove that detectors with imperfect precisions offer limited usefulness. Then we focus on detectors with perfect precision, and we conduct a comprehensive complexity analysis of this optimization problem, showing NP-completeness and designing an FPTAS (Fully Polynomial-Time Approximation Scheme). On the practical side, we provide a greedy algorithm, whose performance is shown to be close to the optimal for a realistic set of evaluation scenarios. Extensive simulations illustrate the usefulness of detectors with false negatives, which are available at a lower cost than the guaranteed detectors.

Throughput and delay analysis of directional‐location aided routing protocol for vehicular ad hoc networks

Summary Vehicular ad hoc networks is an integral component of intelligent transportation systems and it is an important requisite for smarter cities. Network formation and deformation among the vehicles are very frequent because of the variation in speed. Furthermore, for safety applications, messages should not face any kind of delay or collision. Therefore, establishing communication between the vehicles becomes even more challenging. Position-based routing protocols work productively in vehicular ad hoc networks. Only finding an efficient routing protocol does not solve our purport. We need to carefully examine the effect of media access control layer parameters additionally. In the event of collisions, a large number of nodes would be re-transmitting rather than sending fresh packets. A node busy in sending the retransmitted packet is called a backlog node. With an increase in the number of collisions, number of backlog nodes also increases, which affects the delay and throughput. In this article, we present the mathematical modeling of delay and throughput with IEEE 802.11 distributed coordination function (at media access control layer) for directional-location aided routing (D-LAR) position based routing protocol. For performance evaluation, simulation has been done in realistic environment created with SUMO (traffic simulator) and NS-2 (network simulator). Simulation results show the comparison between D-LAR and location aided routing (LAR) on various metrics in terms of delay, packet delivery ratio, routing overhead, throughput, and collision probability. To validate the mathematical model, analytical results has been compared with simulation results. The results confirm that performance of D-LAR is better than LAR in terms of increasing the throughput and reduction in routing overhead and delay.

Effective fuzzy-based location-aware routing with adjusting transmission range in MANET

Mobile ad-hoc network is a collection of mobile nodes. Mobility of the nodes leads to dynamic changes in network topology that may complicate the routing functions. Earlier a variety of topology and position-based routing protocols have been proposed for routing process in MANET. The results reveal that the position-based routing protocols are the better choice for routing functions in comparison with topology-based routing protocols. Therefore, our study is motivated to work for position-based routing so it enhances the performance of network. The proposed protocol aims to reduce the possibility of collisions, delay and decrease the overall communication overhead in network. Hence our scheme first, defines a small triangular shaped requested zone broadcast region which is dynamic in nature. Secondly, fuzzy-based forwarding scheme is proposed which combines both distance and directionbased strategies to determine the next forwarding node to achieve better performance with regard to energy consumption, routing overhead and end-to-end delay. The proposed scheme also focuses on the hole problem in the request zone and proposes a self-adjustable transmission range-based method to overcome the problem. The simulation results show that the proposed scheme outperforms than LAR1 and LARDAR schemes.