1 Internet of Things Software Engineering challenges Network of - - PDF document

1
SMART_READER_LITE
LIVE PREVIEW

1 Internet of Things Software Engineering challenges Network of - - PDF document

Outline Motivation Towards the Assessment of Change in Complex Networks Challenges due to change Reliability testing Benjamin Zeiss (Prof. Dr. Jens Grabowski) Risk assessment and usage profiling Software Engineering for


slide-1
SLIDE 1

1

Towards the Assessment of Change in Complex Networks

Benjamin Zeiss (Prof. Dr. Jens Grabowski) Software Engineering for Distributed Systems Group Georg-August University Göttingen, Germany

Towards the Assessment of Change in Complex Networks (Zeiss, Grabowski) 2

Outline

  • Motivation
  • Challenges due to change
  • Reliability testing
  • Risk assessment and usage profiling
  • Summary and outlook

Towards the Assessment of Change in Complex Networks (Zeiss, Grabowski) 3

Motivation

  • Our background:
  • Test methodology and languages for communicating systems
  • Risk assessments in evolving complex software systems
  • Machine learning in quality assurance
  • Grid interoperability
  • Our interests:
  • How do communicating networks evolve?
  • How do changes affect the dynamic communicating systems?

Towards the Assessment of Change in Complex Networks (Zeiss, Grabowski) 4

Concepts

Ubiquitous computing Pervasive computing Sensor networks Ambient intelligence RFID

Internet of Things

Wireless ad-hoc networks

slide-2
SLIDE 2

2

Towards the Assessment of Change in Complex Networks (Zeiss, Grabowski) 5

Internet of Things

  • Network of services integrated into everyday life that
  • are affordable
  • are special-purpose or multi-purpose
  • use wireless (ad-hoc) communication
  • may have a limited range for communication
  • may use simple protocols
  • have often low energy requirements
  • gather information to pass on

Towards the Assessment of Change in Complex Networks (Zeiss, Grabowski) 6

Software Engineering challenges

  • Reliability assessment after changes
  • Risk assessment / change simulation
  • evolution / predict the future
  • Interoperability between services from different vendors
  • Service conformance
  • Transfer experience from our research to the Internet of Things

Towards the Assessment of Change in Complex Networks (Zeiss, Grabowski) 7

Internet of Things - architectures

Direct communication Mediated communication Meshed services, mediated communication Service Internet gateway Server

Internet Internet

Service Internet gateway Server Mediator

Internet

Services Internet gateway Server Mediator

Towards the Assessment of Change in Complex Networks (Zeiss, Grabowski) 8

Reliability, usage, and change risk

  • New services?
  • Lost connectivity?
  • Service removal?
  • Locality of changes?
  • Stability of dynamic routing?
  • "Compatible" services?

Meshed services, mediated communication

Internet

Services Internet gateway Server Mediator

slide-3
SLIDE 3

3

Towards the Assessment of Change in Complex Networks (Zeiss, Grabowski) 9

Reliability testing

Static Testing Dynamic Testing

  • Analysis of
  • Formal specifications
  • Code
  • Logged runtime data
  • Dynamic testing:
  • Replace mediator
  • Replace services
  • Functional testing
  • Load testing
  • Stress testing
  • Runtime verification:
  • Replace services with "fat" reliability

ensuring services

  • Check constraints on the fly

(temporal logic)

Towards the Assessment of Change in Complex Networks (Zeiss, Grabowski) 10

Usage profiling

  • Network of multi-purpose services:
  • What data does the user actually request or use?
  • Reliability enforcement according to actual network use?
  • Need for different services?
  • Collect usage data
  • Possible uses:
  • Validation of predicted usage
  • Improvement of risk predictions

Towards the Assessment of Change in Complex Networks (Zeiss, Grabowski) 11

Risk assessment

  • Building risk simulation models:
  • Data from dynamic testing , usage profiling, specifcations
  • Latencies, communication frequency, etc.
  • Use measured data to build the risk models
  • Simulate and analyze prospective scenarios

Towards the Assessment of Change in Complex Networks (Zeiss, Grabowski) 12

Summary and outlook

  • Risk model predictions are not limited to changes
  • Partners needed!

Dynamic testing Usage profiling Change prediction Risk model Specification artifacts

slide-4
SLIDE 4

4

Towards the Assessment of Change in Complex Networks (Zeiss, Grabowski) 13

Contact

  • Prof. Dr. Jens Grabowski

grabowski@informatik.uni-goettingen.de

  • Web: http://www.swe.informatik.uni-goettingen.de