Table of Contents

Using the Worksheets Effectively
Selecting a Suitable Machine to Support Logging
Sample Forms

Determining Machine Class and Disk Usage

This section shows you the process of determining:

• The recommended machine required to support your Trend Reporting goals. This is determined by the number of devices that the TrafficDirector application is polling.

• How much disk space is required to store the data you configure the TrafficDirector application to log.

Required disk space is primarily dependent on the following factors:

• Number of agents the TrafficDirector application is polling

• Number of domains and what statistics you configure the TrafficDirector application to sample from each agent

• The interval at which the TrafficDirector application polls the agent for this statistical data

• The aging that you configure for the database tables

The following section contains more information:

• Using the Worksheets Effectively

• Selecting a Suitable Machine to Support Logging

• Sample Forms

Using the Worksheets Effectively

The calculations in the worksheets assume the following:

• On Windows platforms, the NT SQL Server is running on the same machine as the TrafficDirector application.

• The slowest link between the TrafficDirector application and the physical device it is polling is at least T1 (1 MBits/sec)

• Max size of the Host table for all domains is 10,000

• Max size of the Conversations table for all domains is 10,000

Selecting a Suitable Machine to Support Logging

To select a suitable machine to support the logging operations you want to configure, follow these steps. This information will also help you determine if you need additional systems as well.

Fill out Form 1 (Table D-1) according to the following information:

• Each agent counts as one device. An agent is a monitored interface supporting full RMON/RMON2. You can view agents by clicking the Agent radio button in the main screen. Each one of the agents is counted as one device. For example, three Ethernet agents, two Token Ring agents, and two non-Frame Relay agents total to seven devices.

• Each switch counts as one device for every 100 active ports or less. A switch is an agent supporting multiple ports with mini-RMON. You can view switches by clicking the Switch radio button in the main screen. Each switch counts as one device for every 100 active ports or less. For example, a switch consisting of 50 ports is equal to one device; a switch consisting of 200 ports is equal to two devices; a switch consisting of 201 ports is equal to three devices.

• Each Frame Relay agent counts as one device for every 25 PVCs or less. A Frame Relay agent supports multiple PVCs with full RMON/RMON2 which act as generalized agents. You can view Frame Relay agents by clicking the Frame Relay radio button in the TrafficDirector main screen. Each Frame Relay agent counts as one device for every 25 PVCs or less. For example, a Frame Relay agent with 24 PVCs is equal to one device; a Frame Relay agent consisting of 100 PVCs is equal to four devices; a Frame Relay agent consisting of 101 PVCs is equal to five devices.

---

Note If the total number of devices from Worksheet 1 is greater than five, Cisco Systems recommends that the machine be dedicated for logging.

Table D-1: Worksheet 1---Number of Devices

Device Type	Number of Devices
Agent
Switch
Frame Relay
Total

---

Step 2 Based on the total number of devices you entered in Worksheet 1 (Table D-1), use the information in Table D-2 to determine the recommended machine configuration.

Table D-2: Recommended Machine Configuration

		The Configuration of the Machine Should Be:
For This Operating System...	And This Many Devices to be Polled...	CPU Class	Main Memory	Swap Space	Number of Processes	Temporary Disk Space
UNIX	1-25	Class 1	256 MB	256 MB	60	120 MB
UNIX	26-50	Class 2	512 MB	512 MB	110	240 MB
Windows	1-25	Class 1	128 MB	256 MB	60	120 MB
Windows	26-50	Class 2	256 MB	512 MB	110	240 MB

Table D-3 lists the Class 1 and Class 2 recommended models.

Table D-3: Class 1 and Class 2 Recommended Models

CPU Class	Hardware Platform	Recommended Model
Class 1	Sun SPARC	Sparc 20, Ultra 1
Class 1	HP 9000 Series 700 or 800	730/100
Class 1	IBM RS/6000	397/43P model 140
Class 1	Windows NT	Intel Pentium 166
Class 2	Sun SPARC	Ultra 2
Class 2	HP 9000 Series 700 or 800 (Limited to 50 processes.)	HP 777
Class 2	IBM RS/6000	397/43P model 140
Class 2	Windows NT	Intel Pentium Pro

Step 3 Fill out Worksheet 2 (Table D-4) according to the frequency that you want to log Stats, Host and Conversations data for each device type. For example, if you configure logging for Agent Statistics every 5 minutes, enter 5 minutes for Agent/Statistics. Similarly, if you are logging Agent TopN Hosts every 15 minutes, enter 15 minutes in the Agents/TopN Hosts field, and so on.

---

Note You can log information for either TopN Hosts/Convs or All Hosts/Convs, but not both. For example, if you are logging TopN Hosts, All Hosts is not applicable. Similarly, if you are logging All Hosts, TopN Hosts is not applicable.

Table D-4: Worksheet 2---Logging Frequency

Device Type	Statistics	TopN Hosts	All Hosts	TopN Conv	All Conv
Agent
Switch
Frame Relay
Frame Relay PVCs
Agent

---

Step 4 Use Table D-5 to correlate the values you specified in Worksheet 2 to the number of intervals per day. For example, if you specified five minutes, the equivalent interval per day value is 288.

Table D-5: Interval Table

Time	Intervals (I) Per Day
5 min	288
15 min	96
30 min	48
1 hr	24
2 hr	12
4 hr	6
8 hr	3
24 hr	1

Step 5 In Worksheet 3 (Table D-6), enter the intervals per day for each category of each device type.

Table D-6: Worksheet 3---Intervals Per Day

Device Type	Statistics	TopN Hosts	All Hosts	TopN Conv	All Conv
Agent
Switch		N/A	N/A	N/A	N/A
Frame Relay
Frame Relay PVCs			N/A		N/A

---

Note You should include all Full Duplex agents in the Frame Relay agent total. You should include WAN and Fast Ethernet FX interfaces in the Frame Relay total in Step 6, but exclude them from the Agent total.

---

Step 6 Calculate the total number of monitored interfaces (RMON agents) for each device type, as follows:

Step 7 Apply the formulas specified in Table D-7 and Table D-8 for each Device-Statistic combination.

Step 8 Enter the results of each calculation under the appropriate heading.

Note the following abbreviations:

I	Intervals per day specified in Form
R	250 bytes (size of database record)
NH	TopN value for Hosts (default is 50; the actual value for your location is in the SQLDB.dvp file within the USR directory)
NC	TopN value for Conversations (default is 50; the actual value for your location is in the SQLDB.dvp file within the USR directory)
D	Number of Domains (increment by 1 if one of those domains is RMON)
P	Number of active switch ports (from Step 6)
PVC	Number of active PVCs (from Step 6)

Table D-7: Formulas to Calculate Daily Disk Usage per Device-Statistic Combination

Device Type	Statistics	TopN Hosts	All Hosts	TopN Conv	All Conv
Agent	I * R * D	I * R * D * NH	I * R * 10000	I * R * D * NC	I * R * 10000
Switch	I * R *2 * P	Does not apply	Does not apply	Does not apply	Does not apply
Frame Relay	I * R * D * 3	I * R * D * NH	I * R * 10000	I * R * D * NC	I * R * 10000
PVCs	I * R * D * 3 * PVC	I * R * D * NH * PVC	I * R * 10000 * PVC	I * R * D * NC * PVC	I * R * 10000 * PVC

Table D-8: Daily Disk Usage Requirements per Device-Statistic Combination

Device Type	Statistic	TopN Hosts	All Hosts	TopN Conv	All Conv	Total Disk Usage Per Device
Agent	(A1)	(A2)	(A3)	(A4)	(A5)	Sum of A1 through A5 = ___MB
Switch	(S1)	N/A	N/A	N/A	N/A	Sum of S1 through S5 = ___MB
Frame Relay	(FR1)	(FR2)	(FR3)	(FR4)	(FR5)	Sum of FR1 through FR5 = ___MB
PVCs	(P1)	(P2)	N/A	(P3)	N/A	Sum of P1 through P5 = ___MB

Step 9 Enter values in Worksheet 4 (Table D-9) to calculate Total Disk Usage per Day.

Table D-9: Worksheet 4---Total Disk Usage Per Day

Device Type	Total Disk Usage Per Device (From Table D-8)	X	Total Number of Monitored Interfaces for Agents/Frame Relay Agents (From Step 6)	=	Total Disk Usage Per Day
Agent
Switch
Frame Relay			(A)
Frame Relay PVC			1

Step 10 Using the Configure Rollup application in the TrafficDirector Admin window, you can configure aging (in days) for both the detail and the summary tables. The aging value is the actual number of days the data will remain in the database before the dbrolld utility begins to delete the oldest rows first to make room for new information. Enter the number of days you are Aging the Detail database tables: ________days.

Step 11 Use the following formula to calculate long-term disk usage:

Total Disk Usage for all Devices * Aging = Total Long-Term Disk Usage Requirement

Table D-10: Worksheet 5---Results

Recommended Machine Class
Disk Space Requirements

Step 12 Do the following to reduce disk space requirements:

• Log less frequently (increase the time between logging intervals).

• If logging all hosts and conversations, revert back to TopN Hosts and TopN Conversations.

• If already logging TopN hosts and TopN conversations, change N from 50 (default) to a lower number such as 10 or 20. You can find these TopN parameters in the sqldb.dvp file.

• Reduce long-term disk usage by using Configure Rollup to increase the aging frequency of the database tables.

Sample Forms

The sample information provided in Tables D-11 through D-15 illustrate how to use the Trend Reporter calculation forms.

The following agents are monitored in the example:

• Two dual Ethernet agents.

• One switch with 100 active ports.

• One Frame Relay agent with 25 active DLCIs.

Consider the following:

• The number of domains logged per agent or PVC is 20.

• Top50 values for hosts and conversations for 20 domains will deliver database records equivalent to 10 domains.

• The Max size of the Host table in all agents is 10000 for all domains.

• The Max size of the Conversations Table in the agent is 10000 for all domains.

• The slowest link between the TrafficDirector application and the physical device being polled is at least T1 (1 MBits/sec).

Table D-11: Form 1---Number of Devices

Device Type	Number of Devices
Agent	4
Switch	1
Frame Relay	1
Total	6

Table D-12: Form 2---Logging Frequency

Device Type	Statistics	TopN Hosts	All Hosts	TopN Conv	All Conv
Agent	5 min	15 min	---	15 min	---
Switch	15 min	---	---		---
Frame Relay	15 min	15 min	---	15 min	---
Frame Relay PVCs	15 min	4 hr	---	4 hr	---

Table D-13: Form 3---Intervals Per Day

Device Type	Statistics	TopN Hosts	All Hosts	TopN Conv	All Conv
Agent	288	96	---	96	---
Switch	96	---	---	---	---
Frame Relay	96	96	---	96	---
Frame Relay PVCs	96	6	---	6	---

The total number of monitored interfaces (RMON agents) for each device type, as follows:

• Total Agents---4

• Total Switches---100

• Total Frame Relay agents---1

• Frame Relay PVCs. Total PVCs---25

Table D-14: Daily Disk Usage Requirements per Device-Statistic Combination

Device Type	Statistic	TopN Hosts	All Hosts	TopN Conv	All Conv	Total Disk Usage Per Device (Sum of)
Agent	(A1) 1.44 Mb	(A2) 12 Mb	(A3)	(A4) 12 Mb	(A5)	A1 through A5 = 25.44 Mb
Switch	(S1) 4.80 Mb	N/A	---	---	---	S1 through S5 = 4.80 Mb
Frame Relay	(FR1) 1.44 Mb	(FR2) 12 Mb	(FR3)	(FR4) 12 Mb	(FR5)	FR1 through FR5 = 25.44 Mb
PVCs	(P1) 36 Mb	(P2) 18.7 Mb	---	(P3) 18.7 Mb	---	P1 through P5 = 73.4 Mb

Table D-15: Form 4---Total Disk Usage Per Day

Device Type	Total Disk Usage Per Device (From Table D-8)	Times Total No. of Monitored Interfaces for Agents/Frame Relay Agents (From Step 6)	Equals Total Disk Usage Per Day
Agent	25.44 Mb	4	101.76 Mb
Switch	4.80 Mb	1	4.80 Mb
Frame Relay	25.44 Mb	1	25.44 Mb
Frame Relay PVC	73.40 Mb	1	73.40 Mb
Total Disk Usage For All Devices Per Day			205.40 Mb

The number of days for Aging the Detail database tables: 31 days

Long-term disk usage:

205.4 Mb * 31 days = 6.367.4 Mb

Recommended Machine Class---Class 1

Disk Space Requirement---6,367.4 Mb

Posted: Mon Feb 8 15:19:24 PST 1999
Copyright 1989-1999©Cisco Systems Inc.