beautypg.com

Non real-time applications, 3 resources estimation for new customers – Kofax Communication Server 9.1.1 User Manual

Page 27

background image

Environment Guide

Version 3.00.04

© Copyright Kofax, Inc. All information is subject to change without notice.

27

The reason for doing so is to be able to test the disk access to this volume using the dedicated

TcDiskTest tool.

If this recommendation is not fulfilled, it will be not possible to test actual disk throughput with the

dedicated TcDisktest tool afterwards

7.) 1 virtual CPU (VCPU) should be used.

8.) All other KCS applications connecting to the TCOSS must be run in separate VM(s). It is recommended

to connect their VNICs (virtual network adapters) to the same vswitch (virtual switch) as the VM of

TCOSS.

9.) In the case of multi-TCOSS instance operation (ASP model), all TCOSS instances may run in the

single VM. In this way storage TCOSS instances communicate with the media server through internal

vswitch and not over the physical wires.

10.) The TC/PerfLog application must be installed along with KCS setup on the TCOSS and TC/FoIP

server) in order to permanently collect key performance indicators as Windows performance counter

log files. See TC/PerfLog Technical Manual for information about the application.

11.) In the case of TC/FoIP it is strongly recommended to use ESX Server 3.02 or later due to following

performance issue with the UDP packets (ESX 3.0x):

http://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=

10172

Non Real-Time Applications

No special requirements.

5.2.3

Resources Estimation for New Customers

In the case of a new customer who would like to deploy the KCS platform directly in their virtualized

environment (without the possibility to measure the real resource needs of their physical servers), refer to

the chapter Resource Estimation and Key Performance Indicators to find initial resource estimates to start

deployment with.

Later, during the operation of the KCS platform, it is required to observe the actual resource utilization. If

the utilization of a particular resource is higher than planned, grant more resources (in the case of CPU or

RAM) and/or

– in the case of disk or network utilization – inform system administrators on increased

utilization so that they could match the new increased requirements with the overall resources distribution

on the particular ESX server.

Refer to the chapter Resources Estimation for Existing Customers for the guide how to measure real

resource needs.