Presentation on theme: "DLCSPM - Canadian Software Support Seminar - April 1998 - DAPSCT1 A means to ensure software supportability Support Analysis for Software R. Somoza"— Presentation transcript:
DLCSPM - Canadian Software Support Seminar - April DAPSCT1 A means to ensure software supportability Support Analysis for Software R. Somoza
DLCSPM - Canadian Software Support Seminar - April DAPSCT2 Presentation Overview n The Support Analysis for Software (SAS) Process Overview n Software Identification and Breakdown n Categorisation of Software n Supportability Analysis n The Support Concept n Interfaces to other disciplines n The SAS database n Reference Documentation
DLCSPM - Canadian Software Support Seminar - April DAPSCT3 Supportability Plan and Case n Software Supportability Plan: As part of the System Supportability Plan, it describes the activities to be undertaken in order to achieve the software supportability objectives. It also describes activities to be undertaken to demostrate achievement of those objectives n Software Supportability Case: A written documenation about how product supportability was verified/developed at each stage of software development as per the SW Supportability Plan It is convenient, in order to ensure software supportability, to frame the management of software supportability around two key components:
DLCSPM - Canadian Software Support Seminar - April DAPSCT4 A Process for Software Supportability n determined n satisfied n demonstrated Any process (based on the Software Supportability Plan) must achieve that the customer requirements for software supportability shall be: Such a process has therefore to n determine those requirements n influence design so that supportability is built into the product n establish the Support Concept and ensure it is implemented n be itself validated
DLCSPM - Canadian Software Support Seminar - April DAPSCT5 The response: n A methodological process that derives supportability requirements on the basis of the support functions to be carried out: Support Analysis for Software
DLCSPM - Canadian Software Support Seminar - April DAPSCT6 What is SAS? Its obvious: nAnA trick of the contractors to ask for more money nBnBoth nSnSome stupid elucubration from an academic that ¬h¬has never done such an analysis, hhas never supported software, ®h®has never talked to the victims of bad support (the poor users), ¯w¯would not even recognize software if it were put on his dinner plate. nAnA combination of all above (remember Murphys Law...) nAnA new geniality of the government to waste even more money (after all, its just the taxpayers who pay...)
DLCSPM - Canadian Software Support Seminar - April DAPSCT7 What is SAS? It is a consistent methodology that seeks the achievement of system and software supportability throughout requirements, specification and design, in order to define the most cost effective support concept that meets the operational requirements, and to ensure that the necessary support infrastructure is in place before the system enters into service.
DLCSPM - Canadian Software Support Seminar - April DAPSCT8 Goals for a SAS Process n Establish supportability requirements in the early programme phases so that they can be properly addressed n Influence design during development in order to ensure software supportability, both for operation and modification n Ensure that supportability problems do not affect negatively the operation and/or the availability of the fielded system n Ensure all required support processes and infrastructure are properly implemented prior to entry into service n Reduce as far as feasible the cost of ownership n Be itself cost-effective (i.e, the savings and/or achieved improvements must outweigh the cost of the SAS process)
DLCSPM - Canadian Software Support Seminar - April DAPSCT9 Structure for a SAS Process Project Life Cycle Pre-Concept Concept Development Post-Development n PROGRAM PLANNING AND CONTROL – Development of an early SAS Strategy – SAS Plan – Program and Design Reviews n MISSION, DEVELOPMENT & SUPPORT SYSTEMS DEFINITION – Use Study – Technological Opportunities – Standarization – Design Factors – Comparative Analysis – Integration of ADP Systems n PREPARATION AND EVALUATION OF ALTERNATIVES – Functional and Non-Functional Requirements – Support System Alternatives – Evaluation of Alternatives & Trade-Offs n DETERMINATION OF SASR REQUIREMENTS – Operational Task Analysis – Software Exception/Problem Support Analysis (FRACAS) – Software Modification Analysis – Software Transition Analysis – Post-Deployment Software Support (PDSS)/Logistics Management Analysis n SOFTWARE SUPPORTABILITY ASSESSMENT (CASE) – Operation – Modification – Problem Reaction – Logistics Management – Lessons Learnd But: It is convenient to continue during the service phase!
DLCSPM - Canadian Software Support Seminar - April DAPSCT10 Outline of a SAS Analysis Software Identification & Breakdown Categorisation Evaluation of Support Alternatives Identification of Support Resources Trade-Offs Selection of Support Option Documentation of Support Concept, CRLCMP Vendor Support Other Support Options Use Study
DLCSPM - Canadian Software Support Seminar - April DAPSCT11 Identification of SW within a Project n The identification/structuring of software within a project has always been a major problem. n One approach that has worked well (e.g., in the EF-2000 project) is the association of software to the hardware where it executes because: – That is the target machine – It is where hardware/software integration testing will be carried out – It is where the software will reside n This strategy can be made consistent with the LSA breakdown n This strategy, however, might not work in networks or modular avionics n BUT: It should be kept in mind that there are TWO different breakdowns of software, depending on the support function to be assessed: – Functional breakdown (i.e, how software is designed) for modification – Physical breakdown (i.e, the loadable elements) for the operation
DLCSPM - Canadian Software Support Seminar - April DAPSCT12 Functional Breakdown Principles n The Functional Software are all those elements that depict the functional/design aspects of interest to the software engineers n This is the Software that has to be considered for Modification Support n It is important that this breakdowns follows as closely as possible the actual software design – Because otherwise it will falsify your support considerations – Because an Audit Trail to the existing design is required – Because that is the starting point for support – Because the software documentation is structured that way n BUT: You can include "Dummy" Software Items to group sets of functionality that are documented as a whole or which have a separate Design, provided that: – Dummy Items are used to provide an additional level of abstraction – Dummy Items do not change the overall structure and functionality of the Design
DLCSPM - Canadian Software Support Seminar - April DAPSCT13 Functional Breakdown of Software Notes: Shaded areas indicate complete designs (CSCIs) CSCI Candidates are framed, Candidates for Integration in double frames, Dummy Candidate Numbers in brackets. S1AA also includes the interfaces to the other CSCIs & acts as if it were the Top-Level Design; S1 S1A S1A A S1AB S1B S1 S1AS1B (S1AA) S1AB S1AAA S1AAB S1AAG S1AACS1AADS1AAE S1AAF S1ABA S1AB B S1ABC S1ABD S1ABE Candidates for Separate Software Support are Framed S1ABE included in design of S1ABA
DLCSPM - Canadian Software Support Seminar - April DAPSCT14 Functional Breakdown Perspectives The Software Breakdown is ONLY one of the possible views of the Software. Support Decisions CANNOT be based ONLY on this perspective. Structure Hierarchical perspective (Functional dependencies) S1AS1B S1 Software Dependencies between the different software items exist both horizontally and vertically. S1A S1B S1 Design perspective (Module coupling & cohesion)
DLCSPM - Canadian Software Support Seminar - April DAPSCT15 Physical Breakdown Principles The Physical Software is all those elements that can be manipulated separately by the operator/user n This is the Software that has to be considered for Operational Support n Two approaches can be used here: – It is part of the hardware element where it is loaded – It is part of the hardware element where it resides n Similarly, several levels of breakdown can be identified: – System – Line-Replaceable Item (Computer) – Shop-Replaceable Item (Computer Card) – Component (Chip)
DLCSPM - Canadian Software Support Seminar - April DAPSCT16 Physical Breakdown of Software L1 Comms System Load L11 Transceiver LRI Load L12 Audio Mgr. LRI Load L11-L Application SW Load L11-F Firmware Load L121-F Audio Firmware Load L12-L Application SW Load L121-S Audio Module SW Load L122-S DVI Module SW Load L122-F DVI Firmware Load LRI SRI Chip Loading level Green indicates groups of loadable elements Other colors indicate individual loadable elements
DLCSPM - Canadian Software Support Seminar - April DAPSCT17 Software Categorisation n A minimum set of information should be collected for Categorisation. n This information should indicate the main parameters that would affect software Support (e.g., Operational Importance, Property Rights, Frequency of Change, etc) n The result of the Categorisation should be a decision to: ã Not continue with the Analysis ã Collect a minimum Data Set ã Explore all possible Support Alternatives Categorisation criteria are important! Too much analysis (expensive) Too little analysis (lousy support) Note that Categorisation can be made on: - Functional Units - Physical Units - Both
DLCSPM - Canadian Software Support Seminar - April DAPSCT18 Selection of SAS Candidates n SAS Candidates are elements which are going to be subject to some kind of supportability analysis n SAS Candidates should be selected on the basis of their operational or supportability significance n Three types of distinct SAS Candidates should be selected: ý Functional Software Elements ý Physical Software Elements ý Data n Each of these has its own characteristics and different supportability problems
DLCSPM - Canadian Software Support Seminar - April DAPSCT19 Selection of Functional Candidates n Inherent Candidates: – The Item that includes all Software in the System and/or LRI – All Items (CSCIs) that have a separate Design – All intermediate Items between the CSCIs and the Top-Level CSCI at which Integration is performed n Potential Candidates: All Support Anomalies such as – Software Items requiring special hardware or software tools – Software Items of different risk classes than the software where they are embedded – Proprietary software, such as Run-Time Libraries or COTS Software – Software Items that have different versions for use of the parent software on different platforms (e.g., different I/O handlers, so that the Software can run both on an A/C and on the Flight Simulator) – Different programming languages (specially weird ones) – Deviations to the Design Environment – Reused Software
DLCSPM - Canadian Software Support Seminar - April DAPSCT20 Selection of Physical Candidates n For Software: – All software executables that can be loaded and/or installed separately or that require different loading/installation tasks – All groups of the above that are loaded/installed together in one single operation n For Data: – All databases – All data blocks (e.g., mission data) that: n Are manipulated as a single entity n Require specific software for creation, manipulation or evaluation (e.g., post-flight analysis) n Are installed, loaded or unloaded by means of a specific task
DLCSPM - Canadian Software Support Seminar - April DAPSCT21 Function-oriented Supportability Analysis n It is convenient to carry out the Supportability Analysis from the point of view of the support functions that have to be carried out, and not from the point view of the product n When considering the functions (e.g., modify software, operate software), it is much easier to determine the necessary tasks, as well as those aspects that will simplify those tasks (e.g., software modularity helps modification, installation requires a loading device) n A first assessment of those tasks early in the development cycle will be useful to determine the software supportability requirements and objectives n But: software supportability in only one of the system supportability parameters, and should be always assessed in this context
DLCSPM - Canadian Software Support Seminar - April DAPSCT22 Analysis for Modification Support n The Analysis for Modification Support should be based both on the software to be modified and the tasks to be carried out to make that modification. n The first step should use a set of standard tasks (say, from IEEE 1219, ISO 12207, MIL-STD-498 or equivalent). n The standard tasks can be usually related to a specific hierarchical level within the software. n A cost or effort estimation tool (e.g., COCOMO) can provide the effort for the modification task. On the basis of existing metrics or software engineering studies, this effort can be statistically distributed over the different tasks. n The next step consists of identifying the support elements required to carry out those tasks. The use of those elements usually can be derived from the effort and duration of the tasks.
DLCSPM - Canadian Software Support Seminar - April DAPSCT23 Determination of Effort n A cost-estimation tool is used to determine the effort, manpower and duration for each phase of a typical modification n This is adjusted to the number of annual changes n The resulting value, with minor adjustment, can be considered to be the use of the necessary support resources. Image courtesy of Price Systems
DLCSPM - Canadian Software Support Seminar - April DAPSCT24 Determination of Support Resources n Each Support Task might require specific Support Resources to be carried out n These Resources might require, on their turn, other Resources Program Design Program Source Code Ada Programmer Training in Ada Ada LRM User Manuals Operating System Host Computer Coding Ada Program Ada Compiler
DLCSPM - Canadian Software Support Seminar - April DAPSCT25 Trade-offs n Effort n Cost of Ownership n Response Times Trade-Offs are simulated by creating a Support Environment with all its parameters and then comparing: There are tools that permit to consider different support alternatives and assess the most adequate one: n Operational Impact n Required Investment n Technological interests
DLCSPM - Canadian Software Support Seminar - April DAPSCT26 Analysis for Operational Support n Unfortunately, there are no standards for Operational Support from which a set of standard tasks could be derived n One practical approach is to derive a set of Support Initiators, that is, a set of events that might affect software n These Support Initiators result in a process to be modeled, from which resources can then be identified n It is convenient that this is harmonised with the LSA Process, as many of the Support Initiators can be identified through FMECA. Tools used for LORA can often be also used for this purpose n Typical Support Initiators include: – New software release / new firmware – Repair of computing hardware – Corruption of data or executable code (including virus infection) – System/Software failure – New system mission (with change of software and/or mission data)
DLCSPM - Canadian Software Support Seminar - April DAPSCT27 Operational Support Issues n Supportability decisions might have severe operational impact. Some examples: – Where is software loaded? (increases spares or response time) – Are computers stored with or without software? (configuration control, response times) – How many software versions/variants are at each site? (config.) – How do you ship computers containing classified software for repair? (security problem) – Is the bootstrap loader loadable? (might be fun if power goes away while loading it!) – Do you need to test after loading? (safety issues) – Can you actually shut down the system to install new software? – What do you do if the safety-critical software has a problem?
DLCSPM - Canadian Software Support Seminar - April DAPSCT28 Analysis for Logistics Management n There are no standards either for Logistics Management Support from which a set of standard tasks could be derived n Tasks can be identified by modelling the Logistics Management functions or by using a set of Support Initiators as triggers for the process to be carried out n Typical Logistics Management Support functions include: – Problem reporting and corrective actions – System/software configuration control – Packaging – Distribution – Installation and checkout – User Support (including help desk, technical representatives) n Process modelling tools can be used to determine support effectivity n The classic Integrated Logistic Support (ILS) aspects such as Training, Technical Publications, etc, should be also covered here.
DLCSPM - Canadian Software Support Seminar - April DAPSCT29 Data Support n Data Support is often neglected despite the fact that it often is critical to mission achievement n It is, by its mainly electronic nature, often handled in a similar way as software, with which it has a strong dependency n In this context, it is convenient to analyze its support together with that of the software (though not in the same way) n Support aspects to be considered include: – Data preparation software and hardware – Need for validation – Media, transmission networks, security aspects – Usage (e.g. by other programs, or other computers) – Loading/unloading process and/or tools – Compatibility aspects (e.g., with software, or certain system configurations) – Size, data formats
DLCSPM - Canadian Software Support Seminar - April DAPSCT30 Establishment of the Support Concept, CRLCMP n The Support Concept provides the global view about how the Software Support will addressed n It is coherent - Activities related to different support functions are grouped into processes, which themselves are grouped in accordance with the location (level) and the people (agents) involved, for each software product and all resources and facilities that are available for support purposes. n This overall concept is then described in the Computer Resources Life-Cycle Management Plan (CRLCMP) Support Classes Support Profile Support Functions O p e r a t i o n L o g i s t i c s M a n a g m e n t M o d i f i c a t i o n P r o c e s s e s P r o d u c t E n v i r o n m e n t Levels Agents Activities
DLCSPM - Canadian Software Support Seminar - April DAPSCT31 Supportability CASE n The Supportability CASE is demonstrated by: – Quantitative Evidence – Qualitative Evidence – Historical or Comparative Evidence n This demonstration can be also achieved by means of: – The reports generated from the SASR – The CRLCMP
DLCSPM - Canadian Software Support Seminar - April DAPSCT32 Interfaces to other disciplines n SAS has also a link to the classic ILS disciplines, to which it issues requirements for Training, Support Equipment and Support Software, Technical Publications, Facilities, etc., in a similar way as LSA n The Link to LSA is established through the Physical Software breakdown and its relationship with the hardware n It is convenient that LSA (if carried out) documents the SAS results for the Operational Support, so that all operational matters are stored together n The link to RM&T is also only for operational purposes - RM&T for Software Modification is carried out as part of the SAS Analysis n The link to Engineering is through the software design. It is convenient that SAS participates as part of an Integrated Project Team (IPT), or at least as a reviewer during the whole software design.
DLCSPM - Canadian Software Support Seminar - April DAPSCT33 Interface to LSA The LSAR and SASR should be linked, in order to obtain overall system consistency and ensure system supportability The link are the loadable elements and their associated hardware LSA and SAS have to collaborate in this context, so as to ensure that no incompatible support decisions are taken: (e.g., incompatible loading levels after HW repair and after SW update)
DLCSPM - Canadian Software Support Seminar - April DAPSCT34 The SAS Database (SASR) n It provides a single master data repository for all software related information (everybody works on the basis of the same data) n It integrates such information from multiples sources, for all software in one single system or across multiple related systems n It provides an overview of all required support resources, thus making it easier to identify commonalities or consider combined support n It can be queried to seek out specific information, such as software metrics, system load, deviations.... n It can be linked to software engineering tools or information, as well as to logistics-related information (the LSAR) Though often disconsidered in the Supportability Analysis, the use of a database for all software supportability information is of great importance since: However: There is no standard for such a database and the LSAR cannot be used for this purpose
DLCSPM - Canadian Software Support Seminar - April DAPSCT35 Simplified Relational Graph of SASR Support Tasks Training Equipment Software Facilities Publications Staff Media Host Hardware Software-Specific Information Support Resources
DLCSPM - Canadian Software Support Seminar - April DAPSCT36 Conclusions n A Supportability Analysis can be carried out in a similar way as the LSA Process (but dont try to use MIL-STD-1388!) n It is not specially difficult, but it is convenient that it is done by a mix of software engineers and logisticians n Critical to this is the establishment of a software Supportability Plan in the early programme phases n Not only support can benefit from this approach, even the design process might be optimized because of it n The Supportability Case prevents that supportability is killed due to schedule or cost constraints - it is a contractual requirement! But it should be always kept in mind: n Supportability for its own sake is a waste of money!
DLCSPM - Canadian Software Support Seminar - April DAPSCT37 Some Reference Documentation n DEF STAN 00-60, Integrated Logistic Support, Issue 2, October 1996, Part 3: Logistic Support Analysis Application to Software Aspects of Systems n SAE Standard JA1004 (Draft), Software Supportability Program Standard, in SAE Ballot n MIL-HDBK-347, Mission-Critical Computer Resources Support, May 1990 n SAE Aerospace Information Report AIR5121, Software Supportability - An Overview, January 1997 n SAE Report JA1006 (Draft), Software Support Concept, in SAE Ballot n SAE Recommended Practice JA1005, Software Supportability Implementation Guide (Draft) n Software Logistics Planning Handbook, US Army, CECOM, October 1995 n ISO Standard 12207, Information Technology - Software Life Cycle Processes, August 1995 n IEEE Standard 1219, IEEE Standard for Software Maintenance, June 1993 n MIL-STD-498, Software Development and Documentation, December 1994 n SAE G-11 Software Committtee Homepage (http://www.sae.org/TECHCMTE/g11soft.htm) n DEF STAN Homepage (http://www.demon.co.uk/ilsuk.html)