PON data centre design with AWGR and server based routing

2018 20th International Conference on Transparent Optical Networks (ICTON)

2018

Self Cite

In this paper, we utilize Mixed Integer Linear Programming (MILP) models to determine the impact of link failures on the performance of shuffling operations in MapReduce when different data center network (DCN) topologies are used. For a set of non-fatal single and multi-links failures, the results indicate that different DCNs experience different completion time degradations ranging between 5% and 40%. The best performance under links failures is achieved by a server-centric PON-based DCN.

Section: Methodsmentioning

confidence: 99%

Impact of Link Failures on the Performance of MapReduce in Data Center Networks

Mohamed

2018 20th International Conference on Transparent Optical Networks (ICTON)

2018

Self Cite

“…Recently, a number of novel PON architectures were proposed to manage the inter-rack and intra-rack communication in a data centre [23] - [26]. In [27], [28] we studied one of the PON designs proposed in [25] where Arrayed Waveguide Grating Routers (AWGRs) and servers are used to route traffic. In this paper, we focus on the resilience of the AWGR and server based PON data centre architecture against different link failure scenarios and propose a modified design for improved resilience.…”

Section: Introductionmentioning

confidence: 99%

A Resilient AWGR and Server Based PON Data Centre Architecture

Thabit

2020 22nd International Conference on Transparent Optical Networks (ICTON)

2020

This paper studies the resilience of an AWGR and server based PON DCN architecture against link failure scenarios and proposes a modified design for improved resilience. A MILP model is developed to evaluate the performance of the modified design considering different failure scenarios. The results show a limited increase in power consumption and a large increase in delay under failure scenarios compared to the normal state.

“…Optimization results for the energy efficiency when considering the selection of relaying servers and the provisioning of servers computing resources showed average energy savings of up to 59% compared to random workloads placement. The fourth design which combines the benefits of the switching-centric and the server-center designs is optimized in [25]. Further resources provisioning optimizations were performed in [26] for different PON-based DCNs.…”

Section: !Introductionmentioning

confidence: 99%

Energy Efficiency of Server-Centric PON Data Center Architecture for Fog Computing

Mohamed

2018 20th International Conference on Transparent Optical Networks (ICTON)

2018

In this paper, we utilize Mixed Integer Linear Programming (MILP) models to compare the energy efficiency and performance of a server-centric Passive Optical Networks (PON)-based data centers design with different data centers networking topologies for the use in fog computing. For representative MapReduce workloads, completion time results indicate that the server-centric PON-based design achieves 67% reduction in the energy consumption compared to DCell with equivalent performance.