On Energy Efficiency of Networks for Composable Datacentre Infrastructures

@article{Ajibola2018OnEE,
  title={On Energy Efficiency of Networks for Composable Datacentre Infrastructures},
  author={Opeyemi O. Ajibola and Taisir E. H. El-Gorashi and Jaafar Mohamed Hashim Elmirghani},
  journal={2018 20th International Conference on Transparent Optical Networks (ICTON)},
  year={2018},
  pages={1-5}
}
This paper evaluates the optimal scale of datacentre (DC) resource disaggregation for composable DC infrastructures and investigates the impact of present day silicon photonics technologies on the energy efficiency of different composable DC infrastructures. We formulated a mixed integer linear programming (MILP) model to this end. Our results show that present day silicon photonics technologies enable better network energy efficiency for rack-scale composable DCs compared to pod-scale… 

Figures and Tables from this paper

Energy Efficient Placement of Workloads in Composable Data Center Networks
TLDR
A heuristic for energy efficient placement of workloads in composable DCs is proposed which replicates the trends produced by the MILP model formulated in this paper.
A Network Topology for Composable Infrastructures
TLDR
A mixed integer linear programming (MILP) model that dynamically creates logical networks to support intra logical host communication over the physical network topology is formulated.
Optimizing Resource Availability in Composable Data Center Infrastructures
TLDR
This paper model a theorical problem of resource allocation in a composable data center infrastructure as a bounded multidimensional knapsack and then apply multi-objective optimization algorithms, Non-dominated Sorting Genetic Algorithm (NSGA-II) and Generalized Differential Evolution (GDE3), to allocate resources efficiently.
Performance Analysis of Deep Learning Workloads on a Composable System
TLDR
The design of an enterprise composable infrastructure that is implemented is described and the impact of resource dis-aggregation on representative deep learning benchmarks is evaluated.
A Survey of Big Data Machine Learning Applications Optimization in Cloud Data Centers and Networks
TLDR
A summary of the characteristics of various big data programming models and applications are presented and a review of cloud computing infrastructures, and related technologies such as virtualization, and software-defined networking that increasingly support big data systems are provided.

References

SHOWING 1-10 OF 19 REFERENCES
Energy efficient disaggregated servers for future data centers
TLDR
This paper developed a mixed integer linear programming (MILP) model for energy minimization of the virtual machine (VM) placement problem in data centres implementing DS approach and shows that the average power savings are up to 49% for the different VM types considered.
Future Energy Efficient Data Centers With Disaggregated Servers
TLDR
A mixed integer linear programming (MILP) model is developed to optimize VM allocation for the DS-based data center, including the data center communication fabric power consumption, and an energy efficient resource provisioning heuristic for DS with communication fabric (EERP-DSCF), based on the MILP model insights, is developed.
GreenTouch GreenMeter core network energy-efficiency improvement measures and optimization
TLDR
Energy-efficiency improvements in core networks obtained as a result of work carried out by the GreenTouch consortium over a five-year period are discussed and an experimental demonstration that illustrates the feasibility of energy-efficient content distribution in IP/WDM networks is implemented.
Energy Efficient Virtual Network Embedding for Cloud Networks
TLDR
A heuristic, real-time energy optimized VNE (REOViNE), with power savings approaching those of the EEVNE model, is developed, and the power savings and spectral efficiency benefits that VNE offers in optical orthogonal division multiplexing networks are examined.
Energy efficient survivable IP-over-WDM networks with network coding
TLDR
The results show that implementing network coding can produce savings up to 37% on the ring topology and 23% considering typical topologies, and the impact of varying the demand volumes on the network coding performance is studied.
Distributed Energy Efficient Clouds Over Core Networks
TLDR
A framework for designing energy efficient cloud computing services over non-bypass IP/WDM core networks is introduced and a heuristic for real time VM placement (DEER-VM) that achieves comparable power savings is developed.
Bounds for energy-efficient survivable IP over WDMnetworks with network coding
TLDR
Analytical bounds for the energy efficiency of 1+1 survivable IP over WDM networks using network coding are established, providing verification of the MILP and heuristics proposed previously and an efficient, compact means to evaluate network results and allow the performance of large networks to be determined easily.
On the Energy Efficiency of Physical Topology Design for IP Over WDM Networks
TLDR
A mixed integer linear programming model is developed to optimize the physical topology of IP over WDM networks with the objective of minimizing the network total power consumption and the results show that optimizing the physicalTopology increases the utilization of the renewable energy sources.
Energy efficient virtual machines placement in IP over WDM networks
TLDR
This paper investigates the optimization ofVM placement in IP over WDM core networks considering a VM workload that varies with the number of users served by the VM, and shows that the optimal VM placement in distributed clouds yields up to 23% total power saving compared to a single cloud.
...
1
2
...