CMDB Building Blocks

Joshua SchullerISM Practice Senior ArchitectUnisys Corporation

Building Blocks for Your CMDB

© 2012 WWRUG Canada Inc. All Rights Reserved 2

Objectives and Results

Objective- In this session we are discussing implementation considerations when

deploying ADDM discovery with the CMDB and ITSM.

Results- Fully populated CMDB that is operational with the other ITSM modules

Skills developed- ADDM Integration- CMDB Administration- Performance Tuning


Enriching your CMDB dataAutomated population of the CMDBStraightening out the Product Catalog with NormalizationReconciliation tips and performance tuning Benefits now that we’re done

Agenda


Data Enrichment – Two Types

•Any data from a non-integrated source

•Manual input or “one-off” bulk load

•Examples include Cost Centers, Contracts, Used By Relationships, Sites, etc

User Managed Data

•Examples include ADDM, BBCA, SCCM/Seamless Data Pump

•Could be agent-less or agent-based

•Data derived from scanning

Electronically Discovered

Data


Data Enrichment

User Managed Data- Finding the balance between too much or not enough - Appropriate for KEY pieces of data like Status, Work Info- Will always be a need for some kind of bulk loading capability- Consider a procedure to import a User Managed Data template

spreadsheet into a Staging Dataset with a Reconciliation Job- Additional data load tools may help (e.g. Atrium Integrator or

Effective Technologies Transport)


Data Enrichment

Electronically Discovered Data- Automation is preferred (obviously)- Default ADDM CMDB mapping is a good start- ADDM mappings accomplished via TPL (“The Pattern

Language”) formatted files that can easily be extended- Additionally, ADDM 8.3 contains new features for

Collaborative Application Mapping


Collaborative Application Mapping (CAM)

What is Collaborative Application Mapping?- A workflow process that enriches discovered data by rolling it

up into higher level Business Applications- A set of ADDM features which support the creation of Dynamic

and Durable Application Models (Derived Application Data)

CAM enriches the discovered data within ADDM for a more complete CMDB and ITSM integration



Prototype• Search Datastore• Decide on Tiers• Create Groups /

Subgroups• Create Notes

Share• Generate PDF Report• Request App Owner feedback

Provide Seed Data• Name of Application• Affiliated Hostnames• Database Instance• Process Usernames

ADDM Admin

does this 2Application Owner

does this 1

ADDM Admin

does this 3



Gather Seed Data (what is required?)- Application Name- Basic breakdown of tiers - Hostnames involved (for datastore searches only, we will avoid

hardcoding any hostname into the model)- JMX or J2EE application names - Location of DB config file, DB instance names- Any service or j2ee domain names- Process usernames



Build the Prototype- Search the ADDM datastore using the seed data- Think about the discovered technologies in terms of tiers - Create groups/subgroups (aka the scratch pad or workspace)- Add notes into the model with any questions for the

Application Owner- Create the Prototype PDF report with notes and visualizations- Email the report to the Application Owner for feedback and

review



Collaborate!- Create the Prototype PDF report with notes and visualizations- Email the report to the Application Owner for feedback and

review- Application Owner reviews and provides feedback on the notes

Application Owner input is used to refine the Model

Wash, Rinse, Repeat.. the CAM process



Agenda


Discovery to CMDB/Asset without Normalization Overview

Reconciliation Engine

Normalization Engine

Product Categorization

Product Catalog

BMC Atrium CMDB

BMC Remedy ITSM

Discovery Tool

Asset Management

BMC.Asset Dataset Asset ClassesImport

Dataset

No product catalog entries are reviewed or created.

No product categorization is created.

Data comes in from the discovery source. 1 The data is reconciled to

existing data and goes into the Production Dataset. 2


Discovery to CMDB/Asset with Normalization Overview

Import Dataset

Reconciliation Engine

Normalization Engine

Product Categorization

Product Catalog

BMC Atrium CMDB

BMC Remedy ITSM

Discovery Tool

Asset Management

BMC.Asset Dataset Asset Classes

Data comes in from the discovery source. 1

Normalization follows rules defined to update the categorization, product name and manufacturer name.

2

The data is reconciled to existing data and goes in to the CMDB as visible from ITSM. 5

If the dataset is ‘trusted’, product catalog records will be added for the company. 3

If product catalog records are added for the company, corresponding Product Categorization is added 4


Normalization Alias


Product Catalog


Reconciliation Flow

Merge

Import Dataset

BMC.Asset Dataset

Identify

<<ReconciliationId

Using defined rules (queries) look in the BMC.Asset dataset and see if a match is found to a record in the Import dataset. If found, set the reconciliation ID, else create a new recon ID.

1

If a new record, create it.

If an existing record, update it base on precedence rules.

2


Identification Activity Example

Identification Rules for Computer Class Sequence

'‘ADDM_IntegrationID' = $ADDM_IntegrationID$ 10

'Domain' = $Domain$ AND 'HostName' = $HostName$ AND 'SerialNumber' = $SerialNumber$

20

'SerialNumber' = $SerialNumber$ 30

IMPORT DATASETTokenID = Domain = company.comHostName = LibertySerial Number = 12345Recon ID

BMC.ASSET DATASETADDM_IntegrationID = Domain = company.comHostName = FreedomSerial Number = 12345Recon ID = 7890

1

IMPORT DATASETADDM_IntegrationID = Domain = company.comHostName = LibertySerial Number = 12345Recon ID = 7890


Merge Example

IMPORT DATASETTokenID = Domain = company.comHostName = LibertySerial Number = 12345Recon ID = 7890

BMC.ASSET DATASETTokenID = Domain = company.comHostName = FreedomSerial Number = 12345Recon ID = 7890

2

BMC.ASSET DATASETTokenID = Domain = company.comHostName = LibertySerial Number = 12345Recon ID = 7890

Precedence Rules

BMC.ASSET 10

IMPORT 100

IMPORT2 50


Enriching your CMDB dataAutomated population of the CMDBStraightening out the Product Catalog with NormalizationReconciliation tips and Performance Tuning Benefits now that we’re done

Agenda


Which Application Performance Monitoring tools are available in your environment?- Perfmon, top- Service Availability monitoring (e.g. BEM, Nagios)- End User Experience Mgmt, Deep App Tier Visibility/Tracing

Establish a Performance Baseline- What do the AR and DB utilization look like during normal operation?- How long does it take to execute the Identification and Merge activities?- How long to run the ADDM CMDB Sync?- What is the Transaction Size? How many CI’s?

Performance Tuning

23


Performance Tuning – A Real Example

DB utilization during the Identification Activity- Test 1 is the

baseline- Test 2 is post

indexing- Test 3 is post

index tuning

24


Performance Tuning – Indexing

Adding just a few indexes showed immediate gains - Reconciliation

execution times- Overall system

performance- End user experience

25


In this example we will look at the ADDM transaction performance from a Connectivity perspective- What happens “on the wire”

Similar questions- What other APM tools are available to help- How long does it take to run a CMDB Sync- What does CMDB Sync look like for a single host

Performance Tuning – ADDM to CMDB Synchronization

26



Wireshark capture of a single host CMDB Sync

Summary of Delays- Total response time

of 70.8 seconds- 41.2 seconds spent

in the CMDB- 1.9 MB of

application data transferred



Application Turns- Application Turn is a

change in direction of the conversation

Application turns can be significant- Each turn incurs a

network delay- Magnify network latency - Increase application

response time

431 turns for a single CI!!



Also many Out of Sequence Packets This can indicate- Many retransmissions- Heavy congestion- Error prone link- Network Problem…



Agenda


Benefits now that we’re done

Enhanced Integration- Automatically derived data is populated into the CMDB- The whole is more than the sum of parts

Increased Usability- Fully populated data models lead to a richer user experience- ITSM requests can be easily related to discovered Computer Systems,

Software Instances, etc.

Performance- Negative performance can be dramatic in multiple areas- Value diminishes quickly if user experience impacted

Builds a foundation to map Applications to Services


Questions?


Wrap-up

Contact Information

CMDB Building Blocks

Technology

Transcript of CMDB Building Blocks