The document discusses the challenges of big data visualization in IT management, emphasizing the increasing complexity of data and the need for new techniques to represent it effectively. It proposes a layered hierarchical visualization approach that simplifies complex environments by allowing users to recognize dependencies more easily. The approach includes semantic zooming and layered overviews to create a clearer map of the IT environment.

1. Serge Mankovski
CA Labs Research Staff Members
Big Data Visualization in IT
Management Environment

2. Problem as we see it
— Data is more complex than ever before
− Three Vs of Big Data
− Mushup of structured, semi structured, and unstructured
− Longer time frames
— and we are eager to use long term data more than ever before
− Big promise of Big Data
— Existing visualization techniques mostly built for less complex data
— Time constraints for interactive visualization remain as before
— It is clear in IT management that new techniques are needed to
accomplish common use cases

3. Problematic Use Cases In IT MGMT
Elements of interest
− Single element
• Server
• Router
• Database
−
Groups
• Service
• Network
• Location
Relationship
• Server A hosts Virtual Machine B
• Application A uses Database B
• Service A is contains Application B
3
− Datacenter
• Multiple datacenters @ 20,000+ servers each with
several hundred items
− Mainframe
• Single LPAR can have 30,000+ items
− Cloud Connected Enterprise
• All of the above plus partially transparent cloud
deployments

4. 4
This problem exists in many places

5. The Dilemma
5
Access to detail
Reduce screen complexity

6. 6
Our approach to resolving the dilemma
— Common Approach
− Show as much as possible
− Use various layouts
− Use overview and zooming
— Our Approach
− Show as little as possible
− Use simple layout
− Use semantic zooming and
layered overview

7. Map of IT Environment
— A map of the user’s workspace, where elements of the IT
environment are assigned to a layered structure that
allows the user to quickly recognize dependencies
between areas of the network
7
Enterprise Asset
Terrain
Layer 1
Layer 2
Dependency

8. Definition of Layers
We define a layer in terms of:
− A set of elements (or aggregations) we want
to visualize on the layer
− The resources these elements share
8

9. Defining Hierarchical Layers
Step 1 – Assign elements to the layer
1. Select subset of elements
we want to visualize
2. Assign elements
to layer
Services layer
9

10. Layers Definition
IT Environment elements mapped to layers
Abstraction
Services
Applications
Systems
Networks
Assets
10

11. Defining Hierarchical Layers
Step 2 – Select related resources
1. Select type of resource
dependencies to visualize
2. Select relevant relations
between layer element
and related resources
Services layer
11

12. Defining Hierarchical Layers
Step 3 – Detect and draw dependencies
1. Identify layer elements
with shared resources
2. Draw layer elements to
represent overlapping of
resources
Services layer
12

13. Visualization Map Creation
Elements overlap for each level
Abstraction
Services
Applications
Systems
Networks
Assets
13

14. A Map of the IT Environment
Semantic navigation of complex environment
Services Applications Systems
AssetsNetworks
14

15. Complete Map of the IT Environment
Visual Summary of the Environment based on Filters
Service 5 is selected
Services
Related items in all other levels are highlighted
Applications Systems
Assets
15

16. Navigation
Service 5
Service 5
OR
Service 4
Service 5
AND
Service 3
16

17. Getting to What is Important
Start with a complex IT
environment
Use simple filters to create an
abstract representation of IT
infrastructure
Open areas of interest in
context using traditional
visualization tools
17

18. Semantic Aspects of Layer Formation
18Innovation by CA Labs Copyright © 2010 CA. All rights reserved. February 21st 2012

19. Original System
19

20. Select Green Elements
Select Blue Elements
Defining Blue-Green Layer by Type of Element
20
Select Red Relationship

21. Naming Surfaces of the Layer
21
APPLICATION
SERVCE
USES
Top Class
Bottom Class
Relationship
Set

22. Equivalence
22 CEWIT 2011 November 3rd, 2011
Equivalency in Top and Bottom
Classes
red = gray
yellow = green
Equivalency in Relationship Set
(yellow = green)

23. 23 CEWIT 2011 November 3rd, 2011
Transitive Closure
3
5 64 7
1 2
98 10
3
54
1 2
98 10
1 -> 4 -> 8 => 1 -> 8
2 -> 4 -> 8 => 2 -> 8
2 -> 5 -> 9 => 2 -> 9
3 -> 5 -> 9 => 3 -> 9
3 -> 6 -> 8 => 3 -> 8
3 -> 7 -> 9 => 3 -> 9
Becomes

24. Storage behind visualization
24Innovation by CA Labs Copyright © 2010 CA. All rights reserved. February 21st 2012

25. Graph Cache Behind the Visualization
Graphic Renderer
Web
Service
Database
A
P
P A
P
P
A
P
P
DB
DB
DB DB
Graph database querying system
3rd party
application
25

26. Graph cache performance gains
26

27. 27
Questions?