The Experts below are selected from a list of 15 Experts worldwide ranked by ideXlab platform
Mekhitarian Araxi - One of the best experts on this subject based on the ideXlab platform.
-
Reducing Software Complexity in a Distributed Publish-Subscribe system using Multicast communication
KTH Skolan för elektroteknik och datavetenskap (EECS), 2018Co-Authors: Mekhitarian AraxiAbstract:Systems of distributed character are increasing in size and becoming more complex. Managing and adapting to constant changes of requirements is a challenge during the entire system development life cycle. While new functionalities are implemented, the software may change in design and may lead to poor software quality and increased system complexity. This thesis focuses on tackling the complexity issue in a distributed Electronic Warfare system used in military aircraft. The system consists of a server and several clients which acts as publishers or subscribers for different events sent in the system. The communication is based on unicast and uses a publish/subscribe pattern for the client nodes to register as publishers or subscribers to the server. The system is created to Handle Message passing in high rate and is sensitive for Message delays. Due to this, the system is dependent on a reliable network structure with a continuous necessity for development. An implementation of a multicast prototype will be replacing the topic-specific unicast communication and the publish/subscribe registration process to the server. The system will be evaluated by a comparison of the old communication version with the new multicast implementation using software metrics. The result is to evaluate if the behavior and functionality of the distributed Electronic Warfare system change.System av distribuerad karaktär ökar i storlek och blir alltmer komplex. Att hantera och anpassa sig till ständiga kravändringar är emellertid en utmaning under hela systemets utvecklingsprocess. Medan nya funktioner implementeras kan mjukvaran ändras i design vilket kan leda till dålig programkvalitet och ökad systemkomplexitet. Denna rapport fokuserar på att hantera komplexiteten i ett distribuerat telekrigföringssystem som används i militära flygplan. Systemet består av en server och flera klienter som publicerar och prenumererar för olika typer av meddelanden som skickas i systemet. Kommunikationen baseras på enkelsändning och använder ett publish/subscribe meddelandemönster där klienterna registrerar sig som publicerare eller prenumeranter till servern. Systemet är skapat för att kunna hantera höga meddelandehastigheter och har låg tolerans för meddelandeförseningar. På grund av detta är tillförlitlighet i nätverksstrukturen ett essentiellt kvalitetsattribut då nätverket är i behov av en ständig utveckling. En prototyp av multisändning kommer att implementeras och ersätta kommunikationen baserad på enkelsändning och publish/subscribe registreringsprocessen till servern. Systemet kommer att utvärderas genom en jämförelse av den gamla kommunikationsversionen av systemet med den nya multicast-implementeringen med hjälp av mjukvarukomplexitetsmätningar. Resultatet är att utvärdera om beteendet och funktionaliteten hos det distribuerade telekrigföringssystemet ändras
Niti Madan - One of the best experts on this subject based on the ideXlab platform.
-
Scalable and reliable communication for hardware transactional memory
2012Co-Authors: Seth H. Pugsley, Naveen Muralimanohar, Manu Awasthi, Rajeev Balasubramonian, Niti MadanAbstract:In a hardware transactional memory system with lazy versioning and lazy conflict detection, the process of transaction commit can emerge as a bottleneck. This is especially true for a large-scale distributed memory system where multiple transactions may attempt to commit simultaneously and coordination is required before allowing commits to proceed in parallel. In this paper, we propose novel algorithms to implement commit that are more scalable in terms of delay and are free of deadlocks/livelocks. We show that these algorithms have similarities with the token cache coherence concept and leverage these similarities to extend the algorithms to Handle Message loss and starvation scenarios. The proposed algorithms improve upon the state-of-the-art by yielding up to a 7X reduction in commit delay and up to a 48X reduction in network Messages for commit. These translate into overall performance improvements of up to 66 % (for synthetic workloads with average transaction length of 200 cycles), 35 % (for average transaction length of 1000 cycles), and 8 % (for average transaction length of 4000 cycles). For a small group of multi-threaded programs with frequent transaction commits, improvements of up to 8 % were observed for a 32-node simulation
Seth H. Pugsley - One of the best experts on this subject based on the ideXlab platform.
-
Scalable and reliable communication for hardware transactional memory
2012Co-Authors: Seth H. Pugsley, Naveen Muralimanohar, Manu Awasthi, Rajeev Balasubramonian, Niti MadanAbstract:In a hardware transactional memory system with lazy versioning and lazy conflict detection, the process of transaction commit can emerge as a bottleneck. This is especially true for a large-scale distributed memory system where multiple transactions may attempt to commit simultaneously and coordination is required before allowing commits to proceed in parallel. In this paper, we propose novel algorithms to implement commit that are more scalable in terms of delay and are free of deadlocks/livelocks. We show that these algorithms have similarities with the token cache coherence concept and leverage these similarities to extend the algorithms to Handle Message loss and starvation scenarios. The proposed algorithms improve upon the state-of-the-art by yielding up to a 7X reduction in commit delay and up to a 48X reduction in network Messages for commit. These translate into overall performance improvements of up to 66 % (for synthetic workloads with average transaction length of 200 cycles), 35 % (for average transaction length of 1000 cycles), and 8 % (for average transaction length of 4000 cycles). For a small group of multi-threaded programs with frequent transaction commits, improvements of up to 8 % were observed for a 32-node simulation
Naveen Muralimanohar - One of the best experts on this subject based on the ideXlab platform.
-
Scalable and reliable communication for hardware transactional memory
2012Co-Authors: Seth H. Pugsley, Naveen Muralimanohar, Manu Awasthi, Rajeev Balasubramonian, Niti MadanAbstract:In a hardware transactional memory system with lazy versioning and lazy conflict detection, the process of transaction commit can emerge as a bottleneck. This is especially true for a large-scale distributed memory system where multiple transactions may attempt to commit simultaneously and coordination is required before allowing commits to proceed in parallel. In this paper, we propose novel algorithms to implement commit that are more scalable in terms of delay and are free of deadlocks/livelocks. We show that these algorithms have similarities with the token cache coherence concept and leverage these similarities to extend the algorithms to Handle Message loss and starvation scenarios. The proposed algorithms improve upon the state-of-the-art by yielding up to a 7X reduction in commit delay and up to a 48X reduction in network Messages for commit. These translate into overall performance improvements of up to 66 % (for synthetic workloads with average transaction length of 200 cycles), 35 % (for average transaction length of 1000 cycles), and 8 % (for average transaction length of 4000 cycles). For a small group of multi-threaded programs with frequent transaction commits, improvements of up to 8 % were observed for a 32-node simulation
Manu Awasthi - One of the best experts on this subject based on the ideXlab platform.
-
Scalable and reliable communication for hardware transactional memory
2012Co-Authors: Seth H. Pugsley, Naveen Muralimanohar, Manu Awasthi, Rajeev Balasubramonian, Niti MadanAbstract:In a hardware transactional memory system with lazy versioning and lazy conflict detection, the process of transaction commit can emerge as a bottleneck. This is especially true for a large-scale distributed memory system where multiple transactions may attempt to commit simultaneously and coordination is required before allowing commits to proceed in parallel. In this paper, we propose novel algorithms to implement commit that are more scalable in terms of delay and are free of deadlocks/livelocks. We show that these algorithms have similarities with the token cache coherence concept and leverage these similarities to extend the algorithms to Handle Message loss and starvation scenarios. The proposed algorithms improve upon the state-of-the-art by yielding up to a 7X reduction in commit delay and up to a 48X reduction in network Messages for commit. These translate into overall performance improvements of up to 66 % (for synthetic workloads with average transaction length of 200 cycles), 35 % (for average transaction length of 1000 cycles), and 8 % (for average transaction length of 4000 cycles). For a small group of multi-threaded programs with frequent transaction commits, improvements of up to 8 % were observed for a 32-node simulation
