SYMIAN: Analysis and Performance Improvement of the IT Incident Management Process Group 5 Presented by Xiao ZHENG Jiaming GUO 09/10/2012.

SYMIAN:Analysis and Performance

Improvement of the IT Incident Management Process

Group 5Presented by Xiao ZHENG

Jiaming GUO09/10/2012

2

OUTLINE

1. Introduction 2. SYMIAN TOOLS SYMIAN Model of IT Support

Organizations Options for Optimizing Performance Architecture and Implementation Experimental Results 3. Conclusion and Future Work

3

INTRODUCTION

WHAT is SYMIAN (SYMulation for Incident Analysis):

A TOOL: to improve the performance of the incident management;

-build model of Incident management process;

-evaluate performance;

-assess likely improvements;

4

INTRODUCTION

ANALYSIS OFTHE INCIDENTMANAGEMENT IN IT SUPPORT ORGANITIONS

5

INTRODUCTION

6

OUTLINE

1. Introduction 2. SYMIAN TOOLS SYMIAN Model of IT Support

Organizations Options for Optimizing Performance Architecture and Implementation Experimental Results 3. Conclusion and Future Work

SYMIAN Model of IT Support Organizations

We can consider each support group gi with i=1......N. Which is modeled as a G/G/si queue.

Each element of the transition matrix T represents the probability that transfer from support group gi to support group gj

The definition of ai and ci

8

The SYMIAN SIMULATE PROCESS

STARTS

OFF-DUTY OPERATOR

NEXT QUEUE

BUSY OPERATORQUEUE

AVAILABLE OPERATOR

FINISH

9

OUTLINE

1. Introduction 2. SYMIAN TOOLS SYMIAN Model of IT Support

Organizations Options for Optimizing Performance Architecture and Implementation Experimental Results 3. Conclusion and Future Work

Options for optimizing performance(1). Removing support groups. Its effect: the new closure vector will be . (where ≠ 1 ) the new transition matrix is (.

(2). Creating support groups.Its effect: the matrix is extended with the row N+1, j) = It will also be extended with the column i, N+1) =(fromi

(3). Merging support groups. It is equivalent to the removal of each group , followed by the creation of a new group. The ratio between the volume of incidents processed at each group need to be considered.

11

(4). Splitting support groups.It is equivalent to the removal of the old group , followed by the addition of two new groups.The incident volume for each of the two new groups should be considered.

(5). Changing staffing levels, work shifts, and incident management policies.The new staffing level should be defined with respect to the previous one .More realistic working shifts : 8-hour-per-day.Management policies can be changed at each support group.

Options for optimizing performance

12

OUTLINE

1. Introduction 2. SYMIAN TOOLS SYMIAN Model of IT Support

Organizations Options for Optimizing Performance Architecture and Implementation Experimental Results 3. Conclusion and Future Work

ARCHITECTURE AND IMPLEMENTATION

9 components:Configuration Interface(CI)User Interface (UI)Configuration manager(CM)Parameter Identification Module(PIM)Simulation Core (SC) has three sub-components: 1)Incident Generator (IG)2)Incident Response

Coordinator(IRC)3)Incident Processor ( IP)Data Collector (DC)Trace Analyzer (TA)Statistics Module (SM)Reporting Module (RM)

14

OUTLINE

1. Introduction 2. SYMIAN TOOLS SYMIAN Model of IT Support

Organizations Options for Optimizing Performance Architecture and Implementation Experimental Results 3. Conclusion and Future Work

EXPERIMENTAL RESULTS A. Model Inference and Validation

EXPERIMENTAL RESULTS

Model Inference and Validation

EXPERIMENTAL RESULTS B. Evaluation of Configuration Changes

1.calculate the bottleneck score for each support group: BSi = ( FIi + FOi ) *RIi

2.SG22 and SG39 are the major performance bottlenecks of the organization

3.optimizing SG22 and SG39 might improve the performance of the BailUsOut IT support organization

18

EXPERIMENTAL RESULTS B. Evaluation of Configuration Changes

In order to improve the organization performance: increasing the operator efficiency and emulating an improvement in operator performance

launched a 40-round simulation to assess: MTTR improvement: 17.18%, MICD improvement: 2.05%

Better to split SG22 and SG 39 rather than merge SG22 and SG39

small changes to the configuration of support groups with a large fan in and fan out can have a relatively large impact on the whole system behavior

19

OUTLINE

1. Introduction 2. SYMIAN TOOLS SYMIAN Model of IT Support

Organizations Options for Optimizing Performance Architecture and Implementation Experimental Results 3. Conclusion and Future Work

20

CONCLUSIONS AND FUTURE WORK

To optimize the performance of large-scale IT support organizations is complex

SYMIAN is tool for the performance optimization of incident management in IT support organizations

In SYMIAN evaluation, open queuing network models could reproduce the behavior of real-life IT support organizations with a very high degree of accuracy, which calls for further study of bringing deeper understanding

21

THE END THANK YOU