8-3
User
Guide

for

the

Cisco

Application

Networking

Manager

5.2
OL-26572-01
Chapter
8






Configuring

Real

Servers

and

Server

Farms
Information
About

Server

Load

Balancing

•
Hash
Layer4—Selects

the

server

using

a

Layer

4

generic

protocol

load-balancing

method.

•
Hash
URL—Selects

the

server
using
a

hash

value

based

on

the

requested

URL.
You
can

specify

a

beginning

pattern

and

an

ending

pattern

to

match

in

the

URL.

Use

this

predictor
method
to

load-balance

cache

servers.

Cache

servers

perform

better

with

the

URL

hash

method
because
you

can

divide

t

he

contents

of

the

caches

evenly

if

the

traffic

is

random

enough.

In

a
redundant
configuration,

the

cache

servers

continue

to

work

even

if

the

active

ACE

switches

over

to
the
standby

ACE.

For

information

about

configuring

redundancy,

see

the

“Configuring

High
Availability”
section

on

pa

ge

13-1 .

•
Least
Bandwidth—Selects

the

server

with

the

least

amount

of
network
traffic

or

a

specified
sampling
period.

Use

this

type

for

server

farms

with

heavy

traffic,

such

as

downloading

video

clips.

•
Least
Connections—Selects

the

server

with

the
fewest
number

of

active

connections

based

on

server
weight.
For

the

least

connection

predictor,

you

can

configure

a

slow-start

mechanism

to

avoid
sending
a

high

rate

of

new

connections

to

servers

that

you

have

just

put

into

service.

•
Least
Loaded—Selects

the

server

with

the

lowest
load
as
determined
by

information

from

SNMP
probes.

•
Response—Selects
the
server
with

the
lowest
response
time

for
a
specific

response-time
measurement.

•
Round
Robin—Selects

the

next

server

in
the
list

of

real

servers

based

on
server

weight

(weighted
roundrobin).
Servers

with

a

higher

weight

value

receive

a

higher

percentage

of

the

connections.

This
is
the

default

predictor.
Note
The
different

hash

predictor

methods

do
not
recognize

the

weight
value

that

you

configure

for
real
servers.
The

ACE

uses

the

weight

that

you

assign

to

real

servers

only

in

the

round-robin

and
least-connections
predictor

methods.
Related
Topics
Configuring
the

Predictor

Method

for

Server

Farms,

page

8-39
Real
Servers
To
provide

services

to

clients,

you

configure

real

servers

on

the

ACE.

Real

servers

can

be

dedicated
physical
servers

or

VMware

virtual

machines

(VMs)

that

you

configure

in

groups

called

server

farms.
Note
VMs
that

you
define
as

real

servers
can

be

VMs

associated

with

a

VMware

vCenter

Server

that

you
import
into

ANM

(see

the

“Importing

VMware

vCenter

Servers”

section

on

page

5-24 )

and

VMs

that

the
ACE
recognizes

when

configured

for

Dynamic

Workload

Scaling

(see

the

“Configuring

Dynamic
Workload
Scaling”

section

on

pa

ge

8-26 ).
Real
servers

provide

client

services

such

as

HTTP

or

XML

content,

website

hosting,

FTP

file

uploads
or
downloads,

redirection

for

web

pages

that

have

moved

to

another

location,

and

so

on.

You

identify
real
servers

with

names

and

characterize

them

with

IP

addresses,

connection

limits,

and

weight

values.
The
ACE

al

so

allows

you

to

configure

backup

servers

in

case

a

se

rver

is

t

aken

out

of

service

for

any
reason.
After
you

create

and

name

a

real

server

on

the

ACE,

you

can

configure

several

parameters,

including
connection
limits,

health

probes,

and

weight.

You

can

assign

a

weight

to

each

real

server

based

on

its
relative
importance

to

other

servers

in

the

server

farm.

The

ACE

uses

the

server

weight

value

for

the