Browse Prior Art Database

METHOD AND APPARATUS FOR ENHANCED SERVICE AVAILABILITY

IP.com Disclosure Number: IPCOM000248844D
Publication Date: 2017-Jan-17
Document File: 6 page(s) / 2M

Publishing Venue

The IP.com Prior Art Database

Related People

Plamen Nedeltchev: AUTHOR [+5]

Abstract

An analytics engine is provided to perform predictive analytics and provide predictive alerts to enhance service availability on various data center components (e.g., databases, hosts, storage, network, etc.). Unsupervised learning methods are used to create smart profiles and generate heuristic models to describe the work of the system. The behavior of the system is predicted. Data collected from various infrastructure sources are leveraged to produce predictive alerts through the analytics engine. Data from management servers, vendor sites, and other data sources are collected and aggregated based on the predefined rules engine.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 53% of the total text.

Copyright 2017 Cisco Systems, Inc. 1

METHOD AND APPARATUS FOR ENHANCED SERVICE AVAILABILITY

AUTHORS: Plamen Nedeltchev

Jag Kahlon Svenky Sadagopan Beniston Thangaraj

Senthil Kumar

CISCO SYSTEMS, INC.

ABSTRACT

An analytics engine is provided to perform predictive analytics and provide

predictive alerts to enhance service availability on various data center components (e.g.,

databases, hosts, storage, network, etc.). Unsupervised learning methods are used to

create smart profiles and generate heuristic models to describe the work of the system.

The behavior of the system is predicted. Data collected from various infrastructure

sources are leveraged to produce predictive alerts through the analytics engine. Data from

management servers, vendor sites, and other data sources are collected and aggregated

based on the predefined rules engine.

DETAILED DESCRIPTION

In today's era of digitization in the information technology (IT) industry, service

availability is a key contributor in driving business outcomes for revenue maximization.

IT service providers often manage a plethora of databases and application services,

leveraging multilayered hardware and software platforms to support them. Despite

implementing various re-active and pro-active monitoring measures, IT service providers

still continue to experience service unavailability due to planned and unplanned

downtime activities. It would be useful to predict these planned outages and avoid the

unplanned outages while maximizing service availability.

Provided herein is a robust analytics engine capable of processing enormous

amounts of data to improve service availability and enhance service quality. The solution

presented herein uses the following segments, modules and consoles:

Copyright 2017 Cisco Systems, Inc. 2

- Administrative Console

- Interface Modules

- Data Store

- Smart Profile Manager

- Service Advisor

- Data Visualization Layer

An example method is provided as follows.

1. Collection agents collect data from various databases, applications, and infrastructure

endpoints. In this example, data is collected from an archive storage.

2. Collection agents collect the archive storage data along with related database

capacity/performance metrics (e.g., storage, input/output operations per second, latency,

central processing unit, memory, application processes, etc.).

3. This data is stored in the data store, where it is aggregated. In an example, the data is

stored and/or aggregated periodically (e.g., hourly, daily, weekly, etc.). The data may be

stored at a different period than the period at which it is aggregated.

4. A smart profile is developed to take advantage of unsupervised learning with input

data from closed predictive incidents in the future.

5. Using input provided by the smart profile, heuristic methods are used to model and

aggregate the data from the data store.

6. Time series and auto.arima() functions may be used to make predictions b...