Designing knowledge plane to optimize leaf and spine data center
Authors:
Mujahid Sultan,
Dodi Imbuido,
Kam Patel,
James MacDonald,
Kumar Ratnam
Abstract:
In the last few decades, data center architecture evolved from the traditional client-server to access-aggregation-core architectures. Recently there is a new shift in the data center architecture due to the increasing need for low latency and high throughput between server-to-server communications, load balancing and, loop-free environment. This new architecture, known as leaf and spine architect…
▽ More
In the last few decades, data center architecture evolved from the traditional client-server to access-aggregation-core architectures. Recently there is a new shift in the data center architecture due to the increasing need for low latency and high throughput between server-to-server communications, load balancing and, loop-free environment. This new architecture, known as leaf and spine architecture, provides low latency and minimum packet loss by enabling the addition and deletion of network nodes on demand. Network nodes can be added or deleted from the network based on network statistics like link speed, packet loss, latency, and throughput.
With the maturity of Open Virtual Switch (OvS) and OpenFlow based Software Defined Network (SDN) controllers, network automation through programmatic extensions has become possible based on network statistics. The separation of the control plane and data plane has enabled automated management of network and Machine Learning (ML) can be applied to learn and optimize the network.
In this publication, we propose the design of an ML-based approach to gather network statistics and build a knowledge plane. We demonstrate that this knowledge plane enables data center optimization using southbound APIs and SDN controllers. We describe the design components of this approach - using a network simulator and show that it can maintain the historical patterns of network statistics to predict future growth or decline. We also provide an open-source software that can be utilized in a leaf and spine data center to provide elastic capacity based on load forecasts.
△ Less
Submitted 17 September, 2020;
originally announced September 2020.
Efficient Switch Architectures for Pre-configured Backup Protection with Sharing in Elastic Optical Networks (EON)
Authors:
Suthaharan Satkunarajah,
Krishanthmohan Ratnam,
Roshan G. Ragel
Abstract:
In this paper, we address the problem of providing survivability in elastic optical networks (EONs). EONs use fine granular frequency slots or flexible grids, when compared to the conventional fixed grid networks and therefore utilize the frequency spectrum efficiently. For providing survivability in EONs, we consider a recently proposed survivability method for conventional fixed grid networks, k…
▽ More
In this paper, we address the problem of providing survivability in elastic optical networks (EONs). EONs use fine granular frequency slots or flexible grids, when compared to the conventional fixed grid networks and therefore utilize the frequency spectrum efficiently. For providing survivability in EONs, we consider a recently proposed survivability method for conventional fixed grid networks, known as pre-configured backup protection with sharing (PBPS), because of its benefits over the traditional survivability approaches such as dedicated and shared protection. In PBPS, backup paths can be pre-configured and at the same time they can share resources. Therefore, both short recovery time and efficient resource usage can be achieved. We find that the existing switch architectures do not support both PBPS and EONs. Specifically, we identify and illustrate that, if a switch architecture is not carefully designed, several key problems/issues might arise in certain scenarios. Such problems include unnecessary resource consumption, inability of using existing free resources, and incapability of sharing backup paths. These problems appear when PBPS is adopted in EONs and they do not arise in fixed grid networks. In this paper, we propose new switch architectures which support both PBPS and EONs. Particularly, we illustrate that, our switch architectures avoid the specific problems/issues mentioned above. Therefore, our switch architectures support using resources more efficiently and reducing blocking of requests.
△ Less
Submitted 24 December, 2014;
originally announced December 2014.
