CS 521

1
CS 521

• Software Engineering Analysis

2
Course Topics

• Measurement
• Basic Concepts
• Measurement in SE
• Empirical Strategies
• Surveys
• Case Studies
• Experiments
• Empiricisms in SE
• Experiment Planning
• Hypothesis Formulation
• Variable Selection
• Subjects
• Design
• Instrumentation
• Validity
• Threats to Validity
• Evaluation

• Cyber-Physical Systems
• Definition
• Specification
• Design
• Analysis
• Testing
• Security and Trustworthiness

3
cyber-physical system (CPS)

• A system featuring a tight combination of, and
  coordination between, the systems computational
  and physical elements.
• Today, a pre-cursor generation of cyber-physical
  systems can be found in areas as diverse as
  aerospace, automotive, chemical processes, civil
  infrastructure, energy, healthcare,
  manufacturing, transportation, entertainment, and
  consumer appliances.
• This generation is often referred to as embedded
  systems. In embedded systems the emphasis tends
  to be more on the computational elements, and
  less on an intense link between the computational
  and physical elements.
• Unlike more traditional embedded systems, a
  full-fledged CPS is typically designed as a
  network of interacting elements instead of as
  standalone devices. The expectation is that in
  the coming years ongoing advances in science and
  engineering will improve the link between
  computational and physical elements, dramatically
  increasing the adaptability, autonomy,
  efficiency, functionality, reliability, safety,
  and usability of cyber-physical systems.

4
Recent developments

• Infrastructure getting modernized
• Ratio of advanced to regular meters 4.7 (FERC,
  2008)
• Island of Malta becomes smart grid island
• Enemalta and Water Services Corp. to conduct
  remote monitoring 250,000 smart meters
• 400,000 population
• 90M expense
• Network to be completed by 2012
• Remote monitoring, meter reading, and real-time
  management of network
• Real time monitoring and smart meters -gt time of
  day pricing
• Xcel Energy Boulder as first Smart Grid City in
  the U.S.
• First fully integrated smart Grid in U.S.
• PGE rolling out several million smart meters in
  N.Cal
• Alliander - Amsterdam green grid city project
• Several 100 households
• Target completion 2012
• Total 1B investment
• Estimated cost 410/household over 15 years for
  installation of smart grid
• Experted emmissions reduction 40 by 2025

5
The Utility Industry is undergoing rapid change -
Google Power Meter

• Source SmartGridNews.com
• Google is announcing Google PowerMeter, which
  will ultimately become an open platform for home
  energy information.
• PowerMeter is currently in internal beta testing.
  About four dozen Google employees have home
  energy monitors to record their power usage (as
  proxies for the smart meters of the future). A
  Home Energy gadget on their iGoogle home pages
  shows them how much energy they are using. The
  gadget tracks historical data and forecasts
  future trends (similar to the displays available
  for some of Googles finance applications).
• Giving Customer information so they can act
  and help with demand response
• The PowerMeter Platform
• Underneath the PowerMeter gadget is an open
  systems platform that Google equates to Google
  Maps, the highly successful geospatial system
  that has become the foundation for thousands of
  applications.
• Although the company uses the Maps comparison,
  PowerMeter may actually have more in common with
  Google Android and Google Health. Android is a
  platform for building mobile phone applications.
  It deals not just with data, but also with
  hardware. In a similar fashion, Google PowerMeter
  will ultimately need to interface with smart
  meters, thermostats and other devices.
• . Intelligent software with real time
  information pushes consumption away from high
  peak load areas

6
Wikipedia Smart Grid

• A smart grid delivers electricity from suppliers
  to consumers using digital technology to save
  energy, reduce cost and increase reliability.
• Such a modernized electricity network is being
  promoted by many governments as a way of
  addressing energy independence or global warming
  issues.
• For example, if smart grid technologies made the
  United States grid 5 more efficient, it would
  equate to eliminating the fuel and greenhouse gas
  emissions from 53 million cars.
• United States Congress to pass legislation that
  included doubling alternative energy production
  in the next three years and building a new
  electricity "smart grid".
• Alternative fuel sources would require a smart
  and flexible grid

7
Open protocol and standards is the way to go

• "(F) OPEN PROTOCOLS AND STANDARDS. The
  Secretary shall require as a condition of
  receiving funding under this subsection that
  demonstration projects utilize open protocols and
  standards (including Internet-based protocols and
  standards) if available and appropriate." (P.30,
  Section 405 A-F).
• Government plays an important role

8
So where are we headed with Smart Grid?

• Long Term - decades
• - What is the model of the Smart Grid?
• Bringing it alive?
• Knows its status - sensors
• Makes smart decisions intelligent decision
  making (decentralized)
• Fixes/modifies/evolves itself like a living
  organism Control
• Changes its topology
• This is the queen of infrastructures
• Short and Medium
• Is flexible quick repair
• Can take in new energy sources such as wind /
  solar / renewables?
• Can we connect to PHEVs?
• Do we have any smart grids today YES!

9
Organize Thought Leadership Forums on Smart Grid
of Future

• Next Forum, June 18, 2009
• Current technical limitations of 100 year old
  electric grid infrastructure in the United States
• Various visions of the Smart Grid from DOE,
  National Labs, how it relates to technologies
  available today
• Technologies adopted by successful
  implementations of Smart Grid across the US and
  abroad
• Open-systems wireless/comm interface software and
  standards based approach
• Advanced wireless, RFID and RF-sensors
  technologies and their convergence with the grid
• California Energy Commission, Defense Energy
  Support CenterElectric Power GroupEPRIGlobal
  Quality Corp.Hughes Network SystemsISGEC
  GroupIsmb - istituto superiore mario
  boellaLanTech, Inc.Motorola, Inc.

San Diego Gas ElectricSempra Energy/The Gas
CompanySouthern California Edison
CompanySouthern Contracting CompanyTechnoCom
CorporationUniversal Devices, IncUniversity of
South CarolinaUtility Consulting Group
10
Previous forum, March 18, 2009

• Electric Power Group
• Qualcomm Ventures
• Capgemini
• Los Angeles Dept Water Power
• BC Hydro
• Lawrence Berkeley National Laboratory
• Sempra Energy/The Gas Company
• NERC Cyber Security CIP Program
• Siemens
• Oracle Corporation
• Next forum, November,2009

11
RESEARCH

• Variable and uncertain sources
• Solar
• Wind
• Variable sinks
• Appliances
• Spatial and Temporal sources
• and sinks
• PHEVs / hybrids
• Demand Response
• Plug and Play
• Open Architecture
• An intelligent network making decisions

12
Infrastructure upgrade challenge but
opportunity

• Electric grid set up about100 years ago
• 157,000 miles of high voltage electric
  transmission lines
• Since 1990, demand has increased 25
• Construction of power plants has decreased 30
• Recent history .
• Wikiepedia -  The energy crisis was
  characterized by a combination of extremely high
  prices and rolling blackouts. Price instability
  and spikes lasted from May 2000 to September
  2001. Due to price controls, utility companies
  were paying more for electricity than they were
  allowed to charge customers, forcing the
  bankruptcy of Pacific Gas and Electric and the
  public bail out of Southern California Edison.
  This led to a shortage in energy and therefore,
  blackouts. Rolling blackouts began in June 2000
  and recurred several times in the following 12
  months.
• 2003 rolling blackout (Cleveland isolation, 55M
  people affected)
• Opportunity to support changing demands of the
  customer via a flexible infrastructure

13
Demand response

• Demand Response Definition (LBL) DR is a set of
  time-dependent activities that reduce or shift
  electricity use to improve electricity grid
  reliability, manage electricity costs, and
  encourage load shifting or shedding when the grid
  is near its capacity or electricity prices are
  high.
• 
  LBL Demand Response 2004
  Test
• FERC - 8 percent of energy consumers in US have
  demand response program
• Potential demand response from all U.S. programs
  41,000 MW, or 5.8 of peak demand.
• Is increase of 3,400 MW from the 2006 estimate
• largest demand response resource contributions
  from Mid-Atlantic, Midwestern and Southeastern
• Ontario Smart Grid Forum - ..providing
  transparent electricity prices to consumers
  together with time-of-use rates can lead to
  consumption reductions that range from five to
  fifteen per cent.

14
A new grid over the next 25-50 yearsData network
Power Network, Is there a parallel?
15
Challenges and research opportunities

• Lack of clear definition on what the Smart Grid
  will or should look like
• Lack of clear articulation from leaders to
  citizens on the benefits and reason for
  investment
• Lack of on interfaces between devices, networks,
  appliances, meters, infrastructure (need for open
  interfaces)
• Lack of acceptance of problems vendors systems
  sometimes talk even when standard interfaces are
  developed
• Economic justification at the unit level (home,
  office, factory) is challenging
• How does one pay for the investment?
• Who pays?
• How does utility charge for it? Utilities are
  highly regulated
• How does community discount it? Concern about
  certain vendors getting additional advantage
• Rate adjustments incremental would be necessary
• All parties to not share the same vision of the
  Smart Grid
• Evolution versus revolution conflict in
  approaches
• Are there appropriate incentives from government
• Regulatory challenges utilities are regulated
• Infrastructure not ready today to turn on the
  switch
• In the Smart Grid of the Future, what becomes of
  utilities (only a pipe? Or have content what
  is the meaning of content in the Smart Grid of
  the Future)?

16
Where does Wireless Technology Come in?

• Does not require large amounts of fixed
  infrastructure
• New generations of technology can easily replace
  older generations without having to remove cables
• Next generation of appliances can be done easily
• Infrastructure itself can be upgraded frequently
  (e.g. 1G -gt 2G -gt 3G -gt 4G)
• Benefits of wireless, variability in performance
  and resource requirement
• Long range / short range
• Low bandwidth / high bandwidth
• Delays in networks constantly reducing
• Much lower investment to start getting benefits
  of Smart Grid

17
The Wireless Internet of Artifacts 2.0 Edge,
middle, core

• Edgeware - edge of the network generates
• Sensor data from increasingly powerful sensing
• devices
• Variable data rates depending on the application
• E.g. temperature-sensing RFIDs on power lines
• Location (GPS or RTLS) on field equipment
• Sensors are talking to decision making software
  which in turn is routing energy in various
  directions much like a router is forwarding data
  packets to the right destination
• Middleware
• Determines what to do with the sensor data, adds
  intelligence, and then executes it
• Gets high level controls from next layer and
  executes on it
• Centralware
• Makes decision on what needs to get done
• Central repository of information

18
The Wireless Internet of Artifacts 2.0

• Filtration
• Where is the filtration Edge/Middle/Core?
• Where do the rules for filtration come from
  Core?
• Edge node is smart and knows at some level what
  to do
• How does one distinguish between the Edge, Middle
  and Core nodes? Why three levels?
• Aggregation
• Two sensor streams (S1 and S2) need to be
  combined into one (e.g. power sensor status in
  combination with temperature and motion status
  can be used to create a single boolean, at what
  level should the stream be discarded and only the
  boolean propagated further?
• Messaging

19
Cyber security in the Smart Grid

• Cyber and Physical Security is important for the
  Smart Grid
• Security of Wireless Devices is a bigger
  challenge than wired devices
• Devices operating on standard wireless interfaces
  would require standardized security protocols
• Existing protocols such as 802.11i, WEP, WPA,
  Public key/Private Key, etc. require systematic
  investigation and eventually security will scale
  out similar to the net i.e. mixed/heterogeneou
• Definition and meaning of security to an
  appliance needs to be researched
• Physical security would involve adding
  motion/video/infrared sensors which would be
  integrated into the architecture of the system

20
Source CNET Grid gets hacked

• Spies from other countries have hacked into the
  United States' electricity grid, leaving traces
  of their activity and raising concerns over the
  security of the U.S. energy infrastructure to
  cyberattacks.
• The Wall Street Journal on Wednesday published a
  report saying that spies sought ways to navigate
  and control the power grid as well as the water
  and sewage infrastructure. It's part of a rising
  number of intrusions, the article said, quoting
  former and current national security officials.
• There have long been concerns over securing the
  power grid and other infrastructure. Those
  security issues are mounting as utilities use
  more Internet-based communications and software
  to control the grid through smart-grid
  technology.
• A report by security firm IOActive last
  month warned that people with 500 worth of
  equipment and the right training could manipulate
  smart meters with embedded communications in
  people's homes to potentially disrupt operation
  of the grid.
• WIRELESS and Security
• - Business case for Wireless relies on
  scalability, upgradeability and cost
• - What is the model of security on the Smart
  Grid? Just like on the net, there will not be a
  single source of attack and so there will not be
  a single source of security
• - What do you protect and where? You will have
  to made decisions based on cost benefit analysis,
  e.g. in the home the security requirements are
  different from the enterprise Denial of
  service, Protocol hacked, firewalls, encryption
• - Benefits - Mobile phones today are secure so
  wireless on the Grid can be made secure
• -

21
Reconfigurable Wireless Interface for Networking
of Sensors (ReWINS) Architecture
-
-
Hardware design of Intelligent sensor and
wireless interface
Fig. 1 Architecture of Intelligent sensor
Interface

• Multiple protocols
• Variable payloads - depending on the level of
  intelligence required by smart appliance
• Existing devices - Works with existing devices
  and open for scaling up
• Multiple sensors temperature, humidity, motion,
  shock, acceleration, gyroscopic, chemical
• Embedded demand response intelligence within
  low-power Atmel processor
• Accept time of day pricing
• Framework for open AMI connects with
  thermostats, meters, appliances, and HANs

22
WINSmartGrid - Reconfigurable Wireless Interface
for Networking of Sensors (ReWINS)
23
WINSmartGrid Technology

• Low Power technology
• Open architecture
• Standards-based hardware adapted to fit the
  problem resulting in lower overall cost
• Wireless infrastructure for monitoring
• Wireless infrastructure for control
• Two-way communication
• Service architecture with layers - Edgeware,
  Middleware and Centralware
• Over the air download for real-time
  reconfigurability with wireless
• Plug-and-Play approach to network installation
• Reconfigurability - The capability of the
  technology to be reconfigurable allows OTA (over
  the air) upgrade of the firmware to be able to
  handle different appliances, applications,
  sensors, controllers, thermostats, smart meters,
  PHEVs.

24
WINSmartGrid Architecture

• Wireless protocols issues for in-home, in-office
  and in-factory
• Zigbee / 6LoPan / Home plug
• WiFi
• Bluetooth
• Rubee
• EPC / RFID
• Protocols for in-field
• Transmission Infrastructure CDMA, GPRS, LTE,
  WiMAX, Broadband over power lines
• Tracking and sensing technology for meters
• Active versus Passive
• UHF/LF/HF/433Mhz
• Data layer architecture issues
• Bandwidth requirement
• Power constraints
• Security Requirements
• Database requirements

25
Characteristics of WINSmartGrid

• Low Power technology
• Standards-based adapted to fit the problem
  resulting in lower overall cost
• Wireless infrastructure for monitoring
• Wireless infrastructure for control
• Service architecture with three layers -
  Edgeware, Middleware and Centralware
• Open architecture for easy integration
• Plug-and-Play approach to architecture
• Reconfigurability - The capability of the
  technology to be reconfigurable allows OTA (over
  the air) upgrade of the firmware to be able to
  handle different devices, applications, sensors,
  controllers, thermostats, etc.

26
Smart Grid in WINSmartHome

• Three layers
• Research issues
• What is the in-home architecture?
• How does the 3 layer model work?
• Which wireless comm protocol will actually work?
• Are current wireless protocols adequate?
• How can security be done and how important is
  security?

27

• WINSmartGrid UI
• Simplicity for consumer use
• Remote access and control
• Open systems and tools for integration

Energy Manager
28
The Smart Grid Research Center In Progress
Work force (physical, social)
Grid (physical)
Technology (cyber and physical)

• Partnership Academia, Utilities, Government
  Labs, Regulators, Industry
• Demo on UCLA Micro Grid
• March 2010
• Develop demand response capability with UCLA
  WINSmartGrid
• Objective
• to determine how demand response is accepted in
  Micro Grids
• to determine what the reduction demand
  response will be
• to determine what the wireless and mobile
  communications infrastructure will look like for
  a scalable micro grid.
• to connect various smart appliances and devices
  on campus with the objective of studying how a
  heterogeneous wireless infrastructure performs
  when scaled up.
• open-systems to allow vendors to create
  plug-and-play sensor-enabled appliances
• PHEV affect on location-centric and

29
Research Thoughts Wireless Internet for Smart
Grid

• Long Term (25 years vision) versus short term (5
  years)
• Europe is ahead of us
• PHEVs will eventually play a very important role
• Major research issues
• Software architecture
• Integration of sensor interface with demand
  response and building energy infrastructure
• Smart Home Architecture
• Control loop in heterogeneous systems
• Plug and play
• Model of cyber with infrastructure
• Security needs to be solved before utilities will
  start to use wireless on a wide scale

30
The Wireless Internet of Artifacts Version 2.0

• Heterogeneous wireless grid with mobile/roaming
  artifacts (objects, ICT devices people)
• Constantly in communication with the
  infrastructure
• Control decisions made at the edge of the network
  (via Edgeware), in the middle (via Middleware) or
  at the core (Centralware)?
• How is work load and intelligence distributed
  between these layers?
• Messaging engine becomes key to transmit control
  data
• Sitting on these networks are layers of I.P.
• What does this protocol look like? Is the
  current I.P. protocol good enough? Should
  high-media content (such as sending video over
  HAN) adopt a different network approach from the
  rest of the network that only sends period sensor
  data? Is Video input a sensor?
• Allow rich content to move rapidly
• Have intelligence
• location-specific media compression, analysis and
  representation
• Time-specific DRM
• Context specific commerce models

31
The Wireless Internet of Artifacts

• Infrastructure
• With advances in technologies such as EVDO,
  WIMax, Zigbee, UWB, Rubee
• Each wireless internet link will provide SLAs
  that data owner can purchase (Google open model)
• Resources within a wireless network SLA would
  include variables such as
• Bandwidth
• Power utilization (sensor data that needs to be
  sent infrequently between two nodes would opt for
  low-power networks such as a zigbee networks)
• Wireless networks that are remote would utilize
  energy harnessing (green circuits) to offer
  lower-cost transmission
• Designing, managing, controlling, using, and
  benefiting from a new genre of wireless internet
  of artifacts provides for interesting
  opportunities in the future. 

32
Measurement

• What is not measurable, make measurable
  Galileo Galilei (1564 1642)
• Suggests that one of the aims of science is to
  find ways to measure attributes of things we are
  interested.
• Measurement lies at the heart of many systems
  that govern our lives.

33

• Measurement process by which numbers or symbols
  are assigned to attributes of entities in the
  real world in such a way as to describe them
  according to clearly defined rules.
• Entity object or an event in the real world
• Attribute is a feature or property of an entity

34
Measurement

• Definition far from clear cut
• Height of person, but what about IQ, or quality
  of a wine?
• Measuring Instrument?
• Margin for error with best instruments
• What scale is appropriate?
• We can say Joe is twice as tall as Fred, but why
  not yesterday was twice as hot
• We can take the average grade for a quiz, but
  what about the mean of the jersey numbers of the
  Seahawks?

35
Measurement

• Measurement is a direct quantification
• Calculation is indirect, we take measurements and
  combine them into a quantified item that reflects
  some attribute we are trying to understand
  (overall score in a decathlon)
• In SE we often want to combine measurements to
  understand the Big Picture when discussing a
  project

36
Measurement in SE

• Software Engineering describes the collection of
  techniques that apply an engineering approach to
  construction and support of software products.
• Activities include
• Managing
• Costing
• Planning
• Modeling
• Analyzing
• Specifying
• Designing
• Implementing
• Testing
• Maintaining
• Continually striving to improve process and
  product

37
Measurement

• Electrical, Mechanical, Civil Engineering rely on
  measurement measure variables, changes in
  behavior, measuring causes and effects EE uses
  instruments to measure voltage, current,
  resistance to design circuits
• Measurement Considered a luxury in SE

38
Measurement in SE

• Fail to set measurable targets, thus cannot tell
  if we met our goals
• User friendly
• Reliable
• Fail to understand components costs
• Cost of design from cost of coding, testing
• Do not predict quality
• Will or product fail
• We allow anecdotal evidence to convince us to try
  yet another revolutionary technology

39
Measurements in SE

• Measurements made infrequently, inconsistently,
  and incompletely.
• Can they be repeated?

40
Measurements in SE

• Managers
• What does each process cost?
• How productive is the staff?
• How good is the code being developed?
• Will the user be satisfied with the products?
• How can we improve?

41
Measurements in SE

• Engineers
• Are the requirements testable?
• Have we found all the faults?
• Have we met our product or process goals?
• What will happen in the future?

42
Scope of SE Metrics

• Cost and effort estimation
• Productivity Measure
• Data Collection
• Quality Models and Measurement
• Reliability Models

43
Exercise

• 1. Explain the roll of measurement in determining
  the best players in your favorite sport.
• 2. How do you begin to measure quality of a
  software product?

44
Software Engineering Technology Infusion at NASA

• (1) understand the difference between technology
  transfer (the adoption of a new method by large
  segments of an industry) as an industry-wide
  phenomenon and the adoption of a new technology
  by an individual organization (called technology
  infusion), and
• (2) does software engineering technology transfer
  differs from other engineering disciplines. While
  there is great interest today in developing
  technology transfer models for industry, it is
  the technology infusion process that actually
  causes changes in the current state of the
  practice.

45
Tech Transfer Problem

• One reason why there is so much interest in the
  diffusion of innovations is because getting a new
  idea adopted, even when it has obvious
  advantages, is often very difficult. There is a
  wide gap in many fields, between what is known
  and what is actually put into use. Many
  innovations require a lengthy period, often of
  some years, from the time when they become
  available to the time when they are widely
  adopted.
• Problem how to speed up the rate of diffusion of
  an innovation

46
Changes

• process improvement involves changes
• Minor replacing one compiler or editor by
  another
• Major changes that affect the entire development
  process (e.g., using Cleanroom software
  development and eliminating much of the unit
  testing phase).

47
Product Adoption
48
Product Adoption

• First few customers are the oddballs or
  eccentrics of society, who adopt a new product.
  
• Following them are the opinion leaders, who
  then givetheir approval to the product. Society
  then follows these opinion leaders, and product
  growth follows rapidly.
• During the mature stage, as the market saturates,
  growth levels off, giving the characteristic
  S-curve.

49
Technology Transfer

• Gatekeepers. Technology transfer follows a
  similar process. One member of an organization,
  often called the gatekeeper, monitors
  technological developments, and chooses those
  that seem appropriate for inclusion in an
  organization hence opens the gate to the new
  technology.
• Because this role is often informal, it may fall
  naturally to the most creative and technically
  astute individual in an organization. Since the
  gatekeeper is aware of technical developments
  outside of the organization, others in the group
  often look towards this person for guidance. This
  person often is known by the name guru or
  similar sounding monikers.

50
Models for Tech Transfer

• People mover model. In this approach, there is
  personal contact between the developer and the
  user of a technology. Typically there is some
  facilitator within the infusing organization that
  knows about the new technology and wishes to
  import it into the new organization (i.e., the
  gatekeeper).
• This method was found to be the most prevalent
  and effective of all technology transfer methods.
• 1. Spontaneous gatekeeper role assumed by
  organization member.
• 2. Assigned gatekeeper role imposed by management
  on some organization member.
• 3. Umbrella gatekeeper role assumed by another
  organization to impose new technology on others.

51
Tech Transfer Models

• Communication model. In this approach, the new
  technology has appeared in print and, as with the
• people mover model, some facilitator discovers
  the technology and wishes to infuse it into the
  new organization. The print mechanism may be
  internal documentation, conference reports or
  journal publications.

52
Tech Transfer Models

• 3. On-the-shelf model. This approach, relatively
  rare, the new technology to be packaged so that
  non-experts can discover it and learn enough
  about it to begin the infusion process. It
  requires sufficient documentation so that others
  can easily pick it up and use it.

53
Tech Transfer

• Vendor model. This last method requires an
  organization to turn over the task to a vendor to
  sell them a new technology. It effectively turns
  the vendor into the agent of the People mover,
  Communication or On-the-shelf model.

54
Tech Transfer Models

• Rule model. This method uses an outside
  organization to impose a new technology on the
  development organization, which then infuses it
  into its own development process.
• There are many examples within the government
  sector of this last technology transfer model.
  The mandating of the Ada language by the
  Department of Defenses Ada Joint Program Office
  for system development,
• the use of the Software Engineering Institutes
  Capability Maturity Model to evaluate developers
  qualifications for a Department of Defense
  contract,
• the similar process of using international
  standard ISO 9000 in Europe, and the use of
  Federal Information Processing Standards (FIPS)
  by the National Institute of Standards and
  Technology (NIST) are all examples of technology
  transfer imposed by an outside agency.

55
Tech Transfer

• Advocates. Fowler and Levine at the Software
  Engineering Institute have been investigating
  technology transition and have identified an
  extension to the gatekeeper model 6. In their
  model, technology transition is a pushpull
  process
• Producer ? Advocate ? Receptor ? Consumer

56
Tech Transfer

• The produce of the technology needs an advocate
  to export the technology outside of the
  development organization, while the consumer
  organization must have receptors agreeable to
  importing the technology.
• In many instances, however, both the advocates
  and receptors are part of the consumer
  organization, and in practice, this reduces to a
  model very much like the gatekeeper.

57
Maturation

• The original concept for the technology appears
  as a published paper or initial prototype
  implementation.
• 2. The implementation of the technology involves
  the further development of the concept by the
  originator or successor organization until a
  stable useful version is created.
• 3. In the understanding stage, other
  organizations experiment, tailor, expand, modify,
  and try to use the technology.
• 4. In the later transition stage, use of the
  technology is further modified and expands across
  the industry.
• 5. The final maturation stage is reached when 70
  of the industry uses the technology.

58
Maturation

• In 1985, Redwine and Riddle 11 published the
  first comprehensive study of software engineering
  technology transfer,
• Maturation what was the length of time required
  for a new concept to move from being a laboratory
  curiosity to general acceptance by industry.
• In their study, they looked at 17 software
  development technologies from the 1960s through
  the early 1980s (e.g., UNIX, spreadsheets,
  object-oriented design, etc.)
• Technologies, once developed, required an
  average of 7.5 years to become widely available

59
Case Studies

• NASA plays the role of consumer organization
  trying to adopt new technologies.
• These technologies were studied by the Software
  Engineering Laboratory (SEL) at Goddard Space
  Flight Center.
• The SEL was organized in 1976 to study flight
  dynamics software, and since that time it has had
  a significant impact on software development
  activities within the Flight Dynamics Branch
  (e.g., measurement, resource estimation, testing,
  process improvement)
• As a brief overview of SEL operations, the SEL
  has collected and archived data on over 125
  software development projects. The data are also
  used to build typical project profiles against
  which ongoing projects can be compared and
  evaluated. The SEL provides managers in this
  environment with tools for monitoring and
  assessing project status.
• Typically there are 6 to 10 projects
  simultaneously in progress in the flight dynamics
  environment. Each project is considered an
  experiment within the SEL, and the goal is to
  extract detailed information to understand the
  process better and to provide guidance to future
  projects.
• Projects range in size from approximately 10K
  lines of source code to 300K to 500K at the high
  end.
• Projects involve from 6 to 15 programmers and
  typically take from 12 to 24 months to complete.
  All software was originally written in FORTRAN,
  but Ada was introduced in the mid-1980s (see
  below), and there is now an increase in C and C
  programming.

60
Case Study

• Use of Ada
• Ada is a language that was developed by the U.S.
  Department of Defense from 1976 until 1983 as a
  common language on which to build complex
  embedded applications. It is a general purpose
  programming language adaptable to any computing
  environment

61
Case Study - Ada

• Use of Ada on flight dynamics projects was first
  considered in 1985.
• Because of Department of Defense interest in the
  language and because of NASA Johnson Space
  Centers decision to use Ada for Space Station
  software, the SEL desired to look at its
  applicability for other NASA applications.
• The initial stimulus for this activity, then,
  could be a mixture of the communication model
  (i.e., papers were written about Ada),
  on-the-shelf model (i.e., Ada products were being
  sold) and to some extent, the rule model (i.e.,
  since Johnson Space Center adopted Ada, there was
  some pressure to do the same elsewhere within
  NASA).

62
Case Study - Ada

• To truly evaluate the appropriateness of Ada
  within the SEL environment, a parallel
  development of an Ada (GRODY) and FORTRAN (GROSS)
  simulator was undertaken.
• GROSS, as the operational product, had higher
  priority and was developed on time. GRODY, as an
  experiment to learn Ada, had a much longer
  development cycle. In addition, since GRODY was
  known by all to be an experiment, the development
  team was not as careful in its design
• However, the experiences of the GRODY team with
  the typical set of requirements NASA used for
  such products led to a greater interest in
  applying object oriented technology instead as a
  model for future NASA requirements and design
  specifications.
• Although the development of this simulator
  continued until early 1988, by early 1987 it was
  decided that the initial project was sufficiently
  successful to continue the investigation of Ada
  on other flight dynamics problems.
• Elapsed time since start of Ada activity was 30
  months.

63
Case Study - Ada

• Transition phase of technology Transfer. Because
  of the poor performance on the GRODY simulator
  and the problems with developing Ada
  requirements, the SEL undertook a second Ada
  pilot project (GOADA) as an experiment.
• Sufficient confidence in Ada by this time to make
  GOADA an operational product,
• In1990, Ada became the language of choice for
  simulators in the Flight Dynamics Division.
  Transition time was another 30 months.

64
Conclusions

• Infusion mechanisms do not address software
  engineering technologies well.
• Quantitative data is crucial for understanding
  software development processes
• Technology infusion is not free.