1.
Standard Web APIs
for Multidisciplinary Collaboration
prostep ivip Symposium 2019
Axel Reichwein, Koneksys
April 10, 2019
1

2.
Axel Reichwein
● Background in aerospace
engineering
● Since PhD, focus on data
integration
● Previously involved in
standardization efforts related to
SysML (Systems Modeling
Language)
● Now focused on using standard
APIs and creating applications on
top of standard APIs
2

3.
Initially: Focus on using data standards
PhD and postdoc focused on
tradeoff studies
Reconfiguring architecture
model to then automatically
sync detailed
discipline-specific models
Applying rule-based design in
engineering
3

4.
Now: Focus on using API standards
PhD and postdoc focused on
tradeoff studies
Reconfiguring architecture
model to then automatically
sync detailed
discipline-specific models
Applying rule-based design in
engineering
4

5.
Koneksys Consulting Services
Connecting data using standards
● Custom software development
for organizations and vendors
● Research projects
5

6.
Status Quo
6

7.
Crosscutting Concerns Across Disciplines
7
Requirements
Engineering
Design Manufacturing Operation
Traceability
Configuration
management
Change
Management
Reuse

8.
From model silos to connected data
8

9.
2 Trends increasing need for connected data
IoT
New feedback loops needed to make
sense of recorded operational data
Autonomy
Explosion of number of test scenarios
Need to link experienced auton.
vehicle behavior (e.g. saved in data
lakes) with test scenarios (e.g. saved in
systems engineering applications) to
assess coverage of test scenarios and
overall vehicle safety
Design
Procurement
Manufacturing
Operation
9

10.
What does it mean to connect data?
Example: Requirement identifier <- link type -> Simulation parameter identifier
Connection is between IDENTIFIERS of data
Analogy: phone call between 2 persons identified by their phone number
10
Artifact ID1 Artifact ID2
Link Type ID1

11.
What does it mean to connect data?
Example: Requirement identifier <- link type -> Simulation parameter identifier
Connection is between IDENTIFIERS of data
Analogy: phone call between 2 persons identified by their phone number
11
Artifact ID1 Artifact ID2
Link Type ID1
Artifact ID2Link Type ID2
Link Type
ID3

12.
Application Programming Interface (API)
An API can specify the
interface to a database,
software, dataset
WebAPI use Hypertext
Transfer Protocol (HTTP)
De facto Web API
standard now: REST APIs
12
Facebook Google Twitter
API1 API2 API3

13.
REST API Example 1
13
Facebook
GET
"https://graph.facebook.com/{your-user-id}
?fields=id,name "
{
"name": "Your Name",
"id": "your-user-id"
}

14.
REST API Example 2
14
Dropbox
POST
https://api.dropboxapi.com/2/users/get_account
--data "{"account_id": "dbid:AAH4f99..."}"
{
"account_id": "dbid:AAH4f99..."
}

15.
Standardizable Data Aspects
● Identifier
● Version
● Global configuration
● Data model
● Machine-readable API description
● Change events
15

16.
Achieving the digital thread is currently impossible!
Full
connectivity
CAD Software Simulation
API1 API2 API3
Different APIs everywhere!
16

17.
Different API = vendor lock-in
$ $
17

18.
Once Upon a Time - Before the Web
Different protocols to access documents on the internet (Gopher, WAIS, etc...)
No connected documents (hierarchical document structure, no hyperlinks)
Not many persons used the internet
Hypertext existed since 1965, 25 years before the invention of the Web
Lack of standards for Hypertext hindered adoption of Hypertext - no compatibility
between different Hypertext systems
18

19.
Lessons learned from the Web
Seeds for innovation: Open standards + open-source
Web not owned by a software vendor
Any document can connect to any other document
Improved knowledge sharing and collaboration
OSLC driven by similar values than World Wide Web
19

20.
Reusable Web Standards
20
Data
Repository
HTTP
?
URL http://../req2

21.
Machine-readable format for links
21
Needed: format to describe graphs
Different standard data models to describe graphs
Currently most used for data on the Web: Resource Description Framework (RDF)
Artifact ID1 Artifact ID2
Link Type ID1
Artifact ID2Link Type ID2
Link Type
ID3

22.
RDF Example: Schema.org
22
Interoperability between search engines and Web
content creators through schema.org

23.
Standard APIs for Connecting Data
23
Requirements Test cases Simulation
API1 API2 API3
Global_Id1
Global_Id2
Global_Id3
Global_Id4
Global_Id5
Global_Id6
Standard
API
Standard
API
Standard
APIHTTP
URLRDF

24.
Open Services for Lifecycle Collaboration (OSLC)
24
From http://docs.oasis-open.org/oslc-core/oslc-core/v3.0/oslc-core-v3.0-part1-overview.html

25.
Applications decoupled from data
25
Requirements Test cases Simulation
API1 API2 API3
Visualization
Search/Query
Reporting
Modular
architecture
Combining
Applications
and data
through
plug-n-play
Standard
API
Standard
API
Standard
API
Machine
Learning
Link
Management
Modeling

26.
Private/public
Data Web
Distributed
Data Silos
Mashup
Application
Example
Google-like
Search
26
Data
Repository 1
Data
Repository 2
Data
Repository 3
RDF Link Link RDF

27.
Discipline-specific vs. Universal Data Management
Currently, we have discipline-specific data management: Simulation Data
Management (SDM), Product Lifecycle Management (PLM), Application Lifecycle
Management (ALM) etc., and NO Data connectivity
Universal Data Management is needed
● Viewing data as a universal asset
● Viewing data as equal
● Viewing data with its context
● Viewing data as a whole
● Using open standards for creating standard APIs
● Creating applications on top of standard APIs
27

28.
What now?
Request better APIs from vendors
Request OSLC APIs from vendors
Adopt OSLC in your new integration
projects
Perform POC projects with OSLC
Help build the OSLC community
28

29.
Thanks and get in touch!
axel.reichwein@koneksys.com