Using Artificial Intelligence to power Service Virtualization

Using Artificial Intelligence to power
Service Virtualization
Anand Kameswaran
DevOps: API Management and Application Development
CA Technologies
Principal, Product Manager
DO3T36T
@AnandKameswaran
#CAWorld

2 © 2015 CA. ALL RIGHTS RESERVED.@CAWORLD #CAWORLD
For Informational Purposes Only
Terms of this Presentation
© 2015 CA. All rights reserved. All trademarks referenced herein belong to their respective companies. The presentation provided at CA
World 2015 is intended for information purposes only and does not form any type of warranty. Some of the specific slides with customer
references relate to customer's specific use and experience of CA products and solutions so actual results may vary.
Certain information in this presentation may outline CA’s general product direction. This presentation shall not serve to (i) affect the rights
and/or obligations of CA or its licensees under any existing or future license agreement or services agreement relating to any CA software
product; or (ii) amend any product documentation or specifications for any CA software product. This presentation is based on current
information and resource allocations as of November 18, 2015, and is subject to change or withdrawal by CA at any time without notice. The
development, release and timing of any features or functionality described in this presentation remain at CA’s sole discretion.
Notwithstanding anything in this presentation to the contrary, upon the general availability of any future CA product release referenced in
this presentation, CA may make such release available to new licensees in the form of a regularly scheduled major product release. Such
release may be made available to licensees of the product who are active subscribers to CA maintenance and support, on a when and if-
available basis. The information in this presentation is not deemed to be incorporated into any contract.

Abstract
Technology chat to review the details of how Artificial Intelligence is being used to
advance even further the power of CA Service Virtualization. Opaque Data
Processing (ODP) allows CA Service Virtualization to virtualize data in sufficient
detail when the format of the requests and responses is not known; eliminating
the need for a new data handler every time you encounter a new message format.
Make sure not to miss this firsthand look at the advances in ODP being showcased
during CA World 2015.
Anand
Kameswaran
CA Technologies
Principal, Product
Manager

Any sufficiently advanced
technology is indistinguishable
from magic.
- Arthur C. Clarke

ODP: True AI with Service Virtualization
 Virtualizes services without requiring any
knowledge or decoding of the service
protocols
 Applies a genome sequence alignment
algorithm, discovers byte-level patterns in
message protocols
 Virtualizes a much wider range of
protocols without requiring a new
data protocol handler
 Speed and accuracy with entropy weighting
and message clustering
Collaboration project between CA Labs, Swinburne University of Technology in Australia
and the CA Service Virtualization engineering and product teams

The Big Problem - Constraints
“I can’t do anything until I have everything…
and I never have everything!”
ESB
!
!
! ! !
System Unavailable
x x x
x
Invalid data Access Fees
Incomplete
Development

How Service Virtualization Helps
“I have everything I need, when I need it!"
ESB
! ! !
System Unavailable Invalid data Access Fees
Incomplete
Development

Service Virtualization: How does it work?
CAPTURE PROCESS MODEL
Data Protocol Detection
Identification of Operation vs.
other data in transaction
Identification of Magic
Strings/Dates
De-Identify…
Record traffic between existing
systems
Create from engineering
specifications
From sources such as log files,
sample data, packet capture and
application insight files
Living, breathing “live” model
Sophisticated, contextual
behavior
Automatic handling for dynamic
properties

Service Virtualization: How does it work?
CAPTURE PROCESS MODEL
Record traffic between existing
systems
Create from engineering
specifications
From sources such as log files,
sample data, packet capture and
application insight files
Living, breathing “live” model
Sophisticated, contextual
behavior
Automatic handling for dynamic
properties
The DPH
Challenge

The DPH Challenge
 Must know the protocol
‒ What if it is a proprietary protocol and the SME is no
longer available?
‒ What if the protocol is new, and there are no SMEs?
‒ What if you have a copybook and it is incomplete?
 Must be able to identify the operation, body and
syntax
 It takes time to add new protocol support

ODP Essentials
 High accuracy on the 4 well known
protocols tested including IMS, LDAP, SOAP
and Twitter (REST)
 Speed and accuracy with Entropy Weighting +
Message Clustering
 The more data a service observes the more
intelligent it becomes
 Virtualizes services without requiring any
knowledge or decoding of the service protocols
 Applies a genome sequence alignment algorithm,
discovers byte-level patterns in message protocols
 Virtualizes a much wider range of protocols
without requiring a new DPH
Collaboration project between
CA Labs, Swinburne University of
Technology in Australia and the
CA Service Virtualization
engineering and product teams
BRINGS TRUE
ARTIFICIAL
INTELLIGENCE (AI)
TO SERVICE
VIRTUALIZATION
HIGH ACCURACY:
99.6 – 100%

ODP Approach
Current
Approach
Opaque Data
Processing
OPAQUE VSE
raw bytes
request
response
CLIENT
machine
learning
match
VSE
Parse format delimit
request
response
CLIENT
matchdecode

ODP Approach – Byte-Level Alignment
 Receive incoming request from client
 Apply byte-level alignment to transactions in recording to find
closest matching request (Needleman-Wunsch algorithm)
 Perform byte-string substitution in matching response
 Send modified response

Entropy Weighting
Increases accuracy of
request matching
Weighs similar strings in a
message higher than
dissimilar strings
Assumption: The
“Type”/Operation of a
request is more important
to match before the
record information itself
For example: AddUser: 123456789012345
“Add User” – 7 bytes/higher priority
“123456789012345” - 15 bytes/lower priority

ODP Approach with Entropy Weighting
Client
Opaque
Response
Generator
Transaction
Library
String
Substitution
Request
Generated
Response
Request
Matching
Response
Matching Response
Modified Response
Needleman-
Wunsch
Match
Client
Opaque
Response
Generator
Transaction
Library
String
Substitution
Request
Generated
Response
Request
Matching
Response
Matching Response
Modified Response
Calculate
Entropy
Weighing
Weighted
Needleman-
Wunsch
Match

ODP Accuracy with Entropy Weights
Protocol
Simple Method
% Correct
Entropy Weights
% Correct Best Entropy Measure
IMS 77.4 100 Richness
CICS 100 100 Shannon/Richness
LDAP ASN.1 94.2 94.6 Shannon
LDAP Text 100 100 Shannon/Richness
Twitter (REST) 99.5 99.6 Shannon
SOAP 100 100 Any

Capture Raw
Transactions
Cluster
Transactions
Playback

Capture Raw Transactions

Cluster Transactions

Playback

When to Apply ODP
 ODP provides a turn key solution to virtualize
protocols otherwise not supported by CA Service
Virtualization
 When you need reasonable data and you do not
need to force specific behavior
 For encrypted messages, decryption needs to occur
before ODP analysis
 ODP currently treats all transactions as stateless
?

Credits
to the team of engineers and researchers for their hard
work and determination in coming up with this innovative
way to address the ever-growing challenge of protocol
support.
Miao Du, Jean-Guy Schneider, Jun Han and John Grundy –
Faculty of Information and Communication Technologies,
Swinburne University of Technology, Australia
Steve Versteeg – Research Staff Member at CA Labs
Robert Williams Fmr Staff Software Engineer
Chris Kraus Sr. Principal Product Manager
Menka Goyal Sr. Principal Software Engineer
Stefana Muller Advisor, Product Management
Dr. Steve Versteeg
CA Labs
Ms. Miao Du
PhD Candidate
Dr. Jean-Guy Schneider
Prof. John Grundy
Prof. Jun Han
Robert Williams
Chris Kraus
Sr. Principal
Product Manager
CA Technologies
Thank you . . .
Menka Goyal
Sr. Principal
Software Engineer
CA Technologies
Stefana Muller
Advisor,
Product Management
CA Technologies

Q & A

Recommended Sessions
SESSION # TITLE DATE/TIME
DO4T25T
Service Virtualization on Demand—Have Everything,
Whenever and Wherever Needed
11/18/2015 at 04:30 pm
DO4T03S Leading Your DevOps Enterprise Journey with Gene Kim 11/19/2015 at 10:30 am
DO3T24S
Accelerate DevTest with Microsoft® Azure™ and Service
Virtualization on Demand
11/19/2015 at 03:00 pm

Must See Demos
Service
Virtualization
DevOps
Theater 3
DevOps Sim
Experience
DevOps
Theater 3
Application
Test
DevOps
Theater 4
Cont. App.
Insight
DevOps
Theater 4

Follow On Conversations At…
Smart Bar
DevOps
Theater 3 & 4
Tech Talks
DevOps
Theater 3 & 4

For More Information
To learn more, please visit:
http://cainc.to/Nv2VOe
CA World ’15

Using Artificial Intelligence to power Service Virtualization

Recommended

Recommended

More Related Content

What's hot

What's hot (20)

Viewers also liked

Viewers also liked (20)

Similar to Using Artificial Intelligence to power Service Virtualization

Similar to Using Artificial Intelligence to power Service Virtualization (20)

More from CA Technologies

More from CA Technologies (20)

Recently uploaded

Recently uploaded (20)

Using Artificial Intelligence to power Service Virtualization