Browse Prior Art Database

METRIC FOR ROUTING ALGORITHMS BASED ON BIG DATA ANALYTICS

IP.com Disclosure Number: IPCOM000243433D
Publication Date: 2015-Sep-21
Document File: 8 page(s) / 222K

Publishing Venue

The IP.com Prior Art Database

Abstract

A metric for routing algorithms is described based on big data analytics. The metric is for control plane and Software Defined Networking (SDN) routing algorithms to select routes with a lowest probability of failure, leading to improved Service Layer Agreements (SLAs) and minimum economic impact to the customer (even though the selected route may not be that of minimum use of resources). The metric is derived from big data analytics, such as automatically obtained from a posteriori data collected from the network and processed to compute the probability of failure of each transmission component of the network (e.g., fiber spans, cross-connection matrices, etc.). The metric can be used during both provisioning new channels/circuits and during restoration, allowing centralized and/or distributed operation. Also, the metric can be used in conjunction with other routing criteria, such as distance, latency, administrative weight, and the like.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 14% of the total text.

Page 01 of 8

METRIC FOR ROUTING ALGORITHMS BASED ON BIG DATA ANALYTICS

ABSTRACT


[0001]A metric for routing algorithms is described based on big data analytics. The metric is for control plane and Software Defined Networking (SDN) routing algorithms to select routes with a lowest probability of failure, leading to improved Service Layer Agreements (SLAs) and minimum economic impact to the customer (even though the selected route may not be that of minimum use of resources). The metric is derived from big data analytics, such as automatically obtained from a posteriori data collected from the network and processed to compute the probability of failure of each transmission component of the network (e.g., fiber spans, cross- connection matrices, etc.). The metric can be used during both provisioning new channels/circuits and during restoration, allowing centralized and/or distributed operation. Also, the metric can be used in conjunction with other routing criteria, such as distance, latency, administrative weight, and the like.

BACKGROUND


[0002]The two most common routing algorithms in networks, such as optical networks, are Dijkstra and Bellman-Ford. These are shortest-path algorithms with metrics based on parameters such as distance, latency, number of hops, load balance, congestion, and the like. The aforementioned metrics are related to static parameters that minimize the use of network resources, but in real networks some spans are more prone to failure (e.g., aerial fibers are less reliable than buried ones, etc.) and should be avoided to route unprotected traffic. These particular features of customer networks are not currently included in these metrics and the implementation of this metric can automatize the process of customization.


[0003]Further, big data is an emerging field where big data is a broad term for data sets so large or complex that traditional data processing applications are inadequate. Challenges include analysis, capture, data curation, search, sharing, storage, transfer, visualization, and information privacy. For example, companies have been using customer analytics to target advertising, resulting in increases in revenues. Enablers for big data are the decreasing costs of storage and increases in computing capability through scaling of facilities that allow processing of large datasets of information with relative ease. For example, Amazon's Kindle can provide feedback about the titles of books that customers read, their frequency of reading, the number of simultaneous books that are being read as well as other parameters allowing Amazon to know which books their customers are going to buy even before their customers know themselves.


[0004]Big data analytics takes large subsets of complex data and develops insights based on analysis thereof. For example, using the Kindle example, profitable information does not come only from the purchase of the products themselves but also from their use. It would be advantageou...