Computational Intelligence - August 2015 - 65

particle swarm optimization perform
time series forecasting [39].
3.1. Professional Golf Circuit

On the professional golf circuit, more
than 20 professional tours bring together
golfers to compete in a series of tournaments. Each tour has a flagship event
where participation and winning are
important for a player to be eligible to
play in a major golf event. For example,
the highest rated golf tour is the Professional Golfers' Association (PGA) tour,
which culminates with The Players
Championship. The Official World Golf
Ranking (OWGR) aggregates points
from all golf events to rank golfers. Only
the highest ranked players by OWGR,
with some exceptions, can play in the
four major events - PGA Championship,
The Open Championship, Masters Tournament, and U.S. Open. Each of the four
majors automatically receives 100
OWGR points to allocate to winners.

3.2. Professional Tennis Circuit

3.3. Predictive Cloud Computing

The Association of Tennis Professionals
(ATP) and the International Tennis Federation (ITF) host or regulate more than
700 tennis tournaments a year that
include the prestigious four Grand
Slams: Wimbledon, Roland Garros, Australian Open, and US Open. Court surface dictates the type of play where
Wimbledon is played on grass, Roland
Garros is on clay, Australian Open is
synthetic hard, while the US Open is
acrylic hard. Each Grand Glam is worth
2,000 points toward the world rankings
and each maintains the largest amount
of prize winnings within tennis. The
next tennis tournament tiers are the
Masters 1000, 500 Series, and 250 Series
worth 1000, 500, and 250 points respectively. Top-seeded players are generally
eligible to play in Grand Slam tournaments, while only a select few are
invited to enter the contest or win a
qualifying round.

A continuously available virtualized environment that hosts our PCC on a
3-active cluster [31] supports such prestigious events described in Sections 3.1
and 3.2. Figure 3 depicts the architecture of the system. Three security zones
divide each of the data centers. The red
zone is the open Internet. The yellow
zone contains server load balancers and
webservers. The green zones host applications servers. In the green zone, the
following components run: InfoSphere
Streams (Streams) processes social and
machine logging in real time; InfoSphere BigInsights (BI DataNode and
NameNode - Hadoop) runs nightly
map reduce jobs; Unstructured Information Management ArchitectureAsynchronous Scaleout (UIMA-AS)
performs feature extraction and tournament simulation; Web Crawler
(Crawler) detects which webpages for
an event are relevant toward specific

Multi-Site Predictive Cloud Architecture
3-Sited
Production
Red Zone

Yellow Zone

viz.

GSLB
GSLB

www.

Provisioned
Web Server

Internet
www logs

Twitter
Social
Media
Feed

Green Zone
Database
Server

On Demand
Router

SLB

Application
Server

www logs

Web
Clients

Green Zone

Shared
Web Server

SLB
SLB

Predictive
Cloud
Visual

2-Sited Predictive
Cloud

Mid-Level
Log Collection
Server

Log
Collection
Server

App Server
(BigEngine)

UIMAAS
Streams
SLB

Realtime
www logs (json)

www logs

Rabbit MQ
BI
Data Nodes
BI NameNode
Crawler

Backup
Namenode

Key
Event Traffic
Event Visuals
Social Feeds
Crawler

Realtime Logs
Historic Logs

Figure 3 Predictive cloud computing architecture.

august 2015 | IEEE ComputatIonal IntEllIgEnCE magazInE

Table of Contents for the Digital Edition of Computational Intelligence - August 2015