Determine How and Why NSP is "Slow"

Asset ID:	1-75-2050927.1
Update Date:	2016-09-28
Keywords:

Solution Type Troubleshooting Sure

Solution 2050927.1 : Determine How and Why NSP is "Slow"

Applies to:

Oracle Communications Performance Intelligence Center (PIC) Software - Version 7.1 and later
Information in this document applies to any platform.

Purpose

This document help to identify if NSP is really slow (out of specifications) and help to troubleshoot it.

First information that needs to be checked:

What is the workstation configuration
Verify the CPU utilization of the NSP server
Which application is slow? What is the workflow used?
Values are needed to qualify what is called "slow"

Troubleshooting Steps

NSP performances

Is Internet Explorer configured like recommended, chapter Browser setting?
Supported browsers, java version and Flash Player versions are described in the release Notes.
No proxy must be used between Internet Explorer and NSP Server
No antivirus must be configured to analyze the NSP HTML pages
Compare performances against specified performances indicated in the planning guide:
- 1 Mb/s per simultaneous user between the NSP server and the user workstation (test by downloading a file from the apache server)
- Less than 250ms round trip delay (50 ms recommended). From workstation to Server and server to mediation.(test with a ping cmd)
  250 ms round trip delay can impact some applications that require a query/response mechanism that normally occurs on a LAN network, however some response time can be higher on a normal LAN.
- 1 Mb/s per simultaneous user between the NSP and the IXP subsystem

Action	Min	Max
ProTrace first page (see note)	Twenty (20) seconds	sixty (60) seconds
ProTrace next page (see note)	ten (10) seconds	thirty (30) seconds
ProAlarm Viewer inital map open	one (1) minute	three (3) minutes
ProAlarm Viewer refresh	ten (10) seconds	thirty (30) seconds
ProPerf initial dashboard open	one (1) minute	three (3) minutes
ProPerf dashboard refresh	ten (10 seconds)	thirty (30) seconds

ProTrace performances are guaranteed with a small NSP (1 box), if no more than 5 IXP subsystems are configured in the system. If more IXP subsystems are connected, this will result in a potential increased responsiveness, up to 120 seconds.

ProTrace application

How many sessions are displayed per page?
Are quick filters used?
Are the following rules fullfiled:
- Use of the fields with index
- Avoid split queries on fields with indexes
  A smaller Split Period will be very effective for the data that repeats itself almost every second of the time period chosen
Each time provide the query, the time period, the session and the explain plan log

ProPerf application

Check the content of the dashboard:
- How many panels?
- Content of the panels?
  - For dashboards showing data on a given period (like charts) it is recommended to limit to 200 points per KPI displayed and a maximum of 10 KPI per chart.
    E.g. for KPI calculated every 30s, the chart duration should be 2 hours maximum (i.e. 60x2x2=240 points)
    For KPI calculated every 15 mn, this duration should be 2 days.

Attachments

This solution has no attachment