You are on page 1of 74

Chair Ecole Polytechnique – Thales « Engineering of Complex Systems »

Elements of complex systems architecture

Daniel Krob January 2009
Elements of complex systems architecture – P. 1

Contents

Introduction 1st topic 2nd topic 3rd topic Systems engineering context Key systems architecture concepts Some advanced topics Conclusion

P. 3

P. 6

P. 24

P. 52

P. ?

Elements of complex systems architecture – P. 2

Contents

Introduction 1st topic 2nd topic 3rd topic Systems engineering context Key systems architecture concepts Some advanced topics Conclusion

P. 3

P. 6

P. 24

P. 52

P. 71

Elements of complex systems architecture – P. 3

their cost. 4 . The term “system” refers both to the industrial object realized through an industrial process and the highest point of view that one can have when dealing with this industrial object. Elements of complex systems architecture – P.Introduction Philosophy System approach is both a way of thinking and a standardized international engineering practice … … whose objective is to master industrial systems complexity in order to optimize their quality. their time to market and their performance.

model. architectural framework. 5 . etc.Introduction Objectives • Why such a course on system architecture fundamentals? The Elements of complex systems architecture course intends to present and to clarify the key systems architecture concepts both in an intuitive and formalized (as well as possible) way.) used in systems architecture − Giving an example of a (pseudo-formal) architectural description language (SysML) − Presenting complexity measures for complex industrial systems − Discussing the key challenges of systems architecture Elements of complex systems architecture – P. abstraction. • What is it about? − Explaining the industrial background of systems architecture − Introducing to the system architecture paradigm − Presenting the key architectural concepts (systems.

6 P. 71 Elements of complex systems architecture – P. 52 P. 6 . 3 P. 24 P.Contents Introduction 1st topic 2nd topic 3rd topic Systems engineering context Key systems architecture concepts Some advanced topics Conclusion P.

16 Elements of complex systems architecture – P. 7 P. 6 P. 7 .Contents 1st topic 1st sub-topic 2nd sub-topic Systems engineering context Systems in practice Systems enginering & architecture P.

Systems in practice Some complex industrial systems Automobile Information system Rocket Air Traffic Management Aircraft Systems on chip Examples of complex industrial systems Elements of complex systems architecture – P. 8 .

Systems in practice Typologies of systems Electro-mechanical systems & embedded systems Technical systems Hardware systems Software systems Industrial systems « Human » processes Information systems « Human » ware systems Construction management Project management system Typical system decomposition of an industrial system Elements of complex systems architecture – P. 9 .

000 state variables 34 principal sub-systems 14 levels of hierarchy The « car » system … Superstructure Key characteristic 1: hierarchical integration Elements of complex systems architecture – P. Renault.. pneus. Balle. 10 . 2004 Carrosserie Ground contact Engine and propulsion Electronic architecture Automobile = 400. enjoliveurs Direction Planche de bord Pédalier Câblage et électronique habitacle Pièces démontables soubassement Essuyage et lavage Echappement Trains avant/arrière Moyens de retenue Vitrages Climatisation Superstructure Habillages intérieurs Ouvrants Eclairage Source: C. Moteur Accessoires extérieurs Sièges Soubassement Refroidissement et circuit à carburant Boîte de vitesses Suspension GMP Freinage Roues.Systems in practice What are the key systems characteristics in practice? (1) Decomposition into sub-sub-systems Instrumentation Mécanismes coffre.. capot.

11 .Systems in practice What are the key systems characteristics in practice? (2) Information flows Key characteristic 2: industrial systems are permanently evolving … Elements of complex systems architecture – P.

MIT. de Weck. 2006 .Systems in practice Example of practical difficulties: redesign of an hardware system (1) Elements of complex systems architecture – P. 12 Source: O.

Systems in practice Example of practical difficulties: redesign of an hardware system (2) Aluminium → Titanium O(100g) → O(10M$) Elements of complex systems architecture – P. 13 Source: O. MIT. 2006 . de Weck.

500 / 13. 14 Number of studied projects: 8. nor the time delays and/or scheduled budget • Failure = project ended before the expected end or never put into operations 1996 1998 2000 2002 2004 Types of projects: • 71 % : development projects 36 %: traditional legacy development 19 %: oriented object legacy development 16 %: mixt strategy (development + software) • 29 % : software integration 4 % : integrated software without modification 13 %: light integrated software parametrization 6 %: assembling of bought components 6 %: heavy integrated sofware parametrization Elements of complex systems architecture – P.Systems in practice Example of pratical difficulties: information systems (1) 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% 1994 Failure Medium Succces The Chaos study of the Standish Group The only long term study on the software failure in the world ! Succces Medium Failure • Success = project ended by respecting the technical agenda without any time / budget overcrossing • Medium = project ended without respecting neither the technical agenda.500 projects – every 2 years since 1994 Project origins: • 45 %: international companies – 35 %: ME – 25 %: SE • 60 %: USA – 25 %: Europe – 15%: rest of the world The information systems situation .

15 • 80 business units worldwide • 3 different core activities • 100.Systems in practice Example of pratical difficulties: information systems (2) • • • • shareholders satisfaction through a worlwide management standard creation of homogeneous management processes at the group level possibility of permanent access to quality consolidated business figures suppression of 50 M€ of management expenses per year Objectives of the project: They are non technical ! Tension growing Launching of the project • • • • Anormal turn overs Budget excesses Business units resistancy Use problems Collaborative elaboration of the real business requirements Business requirements validation Core model redesign Real beginning COMEX reaction Wave 1 Wave 2 Scheduled end of the project Previsional budget: 150 M€ Wave 3 Core model definition Pilot phases launching 2001 2002 2003 2004 2005 The (real) story of an ERP project in a big international company Elements of complex systems architecture – P.000 people in the world .

6 P. 7 P. 16 . 16 Elements of complex systems architecture – P.Contents 1st topic 1st sub-topic 2nd sub-topic Systems engineering context Systems in practice Systems enginering & architecture P.

Systems engineering & architecture Thinking in terms of “systems” (1) Cheaper & faster airplanes alternatives USA & Canada Heathrow airport London New York JFK airport Europe CDG airport Paris A « good » technical solution depends on the considered system ! Elements of complex systems architecture – P. 17 .

Systems engineering & architecture Thinking in terms of “systems” (2) Other airplane companies Air travel duration & price Destination Decision Traveling destinations (as a system) Protests (through inhabitants) Noise Transfer duration Travelers (as a system) Air travel duration & price Choice Air trafic regulations (as a system) Destination Decision Landing permission Luxury [P/D = 400 €/h] Concorde P/D < 100 €/h Other solutions Preparation Concorde system All durations & prices obtained The Concorde case systemic analysis Elements of complex systems architecture – P. 18 Traveler behavior model D = Transfer duration + Air travel duration .

Systems engineering & architecture Systems engineering history Train Military complex systems Aeronautics Boeing Airbus Electronics Beginning of cold war SAGE NTDS Spatial complex systems Complex civil industrial systems Automobile Ford General Motors PSA Renault Mercury Apollo NASA System Engineering book Software industry 1950 Systems engineering emergence 1960 1970 Systems engineering structuring 1980 1990 2000 Progressive diffusion of systems engineering in the civil industrial world • SAGE = Semi-Automatic Ground Environment = 1st American anti-aircraft defense system • NTDS = Navy Tactical Data System = = 1st American naval defense system Elements of complex systems architecture – P. 19 .

20 .Systems engineering & architecture The two sides of system design (1) Mechanical sub-system Electronical sub-system Thermical sub-system Software sub-system System level Focus is put on the global system behavior and on the sub-systems high level & interface specifications Sub-systems levels Sub-systems are analyzed & specified individually and finely Mechanical sub-system Thermical sub-system Electronical sub-system Software sub-system Elements of complex systems architecture – P.

Systems engineering & architecture The two sides of system design (2) Mechanical sub-system Electronical sub-system Technical system to design Thermical sub-system Software sub-system « Human » designing system Mechanics engineering Thermics engineering System architecture Electronics engineering Software engineering Sub-systems engineering is in charge of the homogenous boxes when system architecture is responsible of the heterogeneous arrows Elements of complex systems architecture – P. 21 .

2004 Elements of complex systems architecture – P. Renault. 22 . Balle.Systems engineering & architecture What does it mean in pratice: the example of a “transversal” function Main sub-systems • Carrosserie • Ground contact • Engine and propulsion • Electronic architecture Speed regulation through obstacle detection Source: C.

Systems engineering & architecture Synthesis: architecture versus analysis Paradigm Key principle Perimeter Building blocks Mindset Description mode Working mode Interaction mode Industrial specialist Analytical Exhaustive understanding Homogeneous system Disciplinary knowledge Uniqueness & certainty Detailed representation Assembling Expertise & local Engineer Architectural Global understanding Heterogeneous system Systems & interfaces Diversity & relativity Perceptions & viewpoints Integration Collaborative & global Architect Elements of complex systems architecture – P. 23 .

6 P. 52 P. 24 . 3 P.Contents Introduction 1st topic 2nd topic 3rd topic Systems engineering context Key systems architecture concepts Some advanced topics Conclusion P. 71 Elements of complex systems architecture – P. 24 P.

43 Elements of complex systems architecture – P. 25 P. 33 P. 25 .Contents 2nd topic 1st sub-topic 2nd sub-topic 3rd sub-topic Key systems architecture concepts What is a system? Systems architecture frameworks Systems architecture description P. 24 P.

26 Dataflows (spatial & temporal dimensions) .What is a system? What do we want to model? Integrative complexity Logical interactions C2 system Logical interactions System level Door Propulsion Electromechanical interactions Shell Electromechanical interactions Abstraction Logical interactions Instructions Logiocal interactions Electromechaniqal interactions Subsystem level Driver board Informations Air Electromechanical interactions Shell body Integration Air entrance Module level Air Air Air Curved tube Electromechanical interactions Air filter Electro mechanical interactions Module level etc. Elements of complex systems architecture – P.

What is a system? A behavioral definition for a system Input (X) Flow = XT Internal (q) Output (Y) Flow = YT Y = F(X.q) A system is a continuous partial function on dataflows that may transform an input dataflow (X) into an output dataflow (Y) depending on its internal state (q) Key note: a behavioral definition is mandatory due to the fact that logical behaviors are the only common points between all different types of homogeneous systems at the level of abstraction that we must use from a systems architecture point of view Elements of complex systems architecture – P. 27 .

28 .What is a system? Key operator 1: integration S S1 S2 S3 S4 S5 No synchronization – Data are taken at the rythm of each system when possible The integration operator is defined by a fixed point semantics (Kahn. 1974) Continuity is the only technical property used to prove system stability Elements of complex systems architecture – P.

Couzot) • Abstraction(Concretization(A)) ⊆ A • C ⊆ Concretization(Abstraction(C)) An abstraction is a non (too) destructive idealization of a set of objects Elements of complex systems architecture – P. P. & R.What is a system? Key operator 2: abstraction Concrete set Abstraction Abstract set Circle Concrete objects: coloured 2D-objects Rectangle Abstract objects: shapes Shape Concretization Circle All possible circles Circle Formal definition (Galois correspondance. 29 .

30 .What is a system? Example of a formal system modeling (1) Hardware systems Humanware Father system Baby Users Software system • Electronic toothbrush • Brushing speed programmable regulator Brushing analysis & reporting Mother System The considered system: an electronic toothbrush + its users Elements of complex systems architecture – P.

ON) (0.What is a system? Example of a formal system modeling (2) E = Electricity flow A =Action flow Base (R.D) Decision = Stop ON Decision = Start OFF D = Decision flow D ∈ {Start.D)=Corpse(E.D.D.OFF) F = Electricity flow D = Decision flow Head C=Head(F.D) = Corpse(E.cleaned}T Decision = Use User C = User(D) The full system is obtained by integration from this description Elements of complex systems architecture – P.E) = Base(A. Use.E) R = Reaction flow E = Electricity flow Corpse (F. Stop}T C = Cleaning flow C ∈ {nothing.cleaning. 31 .

D) = Corpse(E.D)=Corpse(E.OFF) D = Decision flow D ∈ {Start. one must use abstraction / concretization operators Elements of complex systems architecture – P.What is a system? Example of a formal system modeling (3) Corpse (F. Stop}T User C = User(D) Abstraction Abstraction I = Instantaneous effort flow I ∈ [0.10]T Button P = Permanent effort flow P ∈ [0.D.10]T Hand zone Concretization consists here in associating an interval of effort flows to a decision To enter into a detailed modeling of the considered system.D.ON) (0. Use. 32 .

25 P. 33 P.Contents 2nd topic 1st sub-topic 2nd sub-topic 3rd sub-topic Key systems architecture concepts What is a system? Systems architecture frameworks Systems architecture description P. 33 . 43 Elements of complex systems architecture – P. 24 P.

Systems architecture frameworks System visions

Viewpoint

Answers to the question

Some associated keywords

Examples (e-toothbrush)

Operational

Why ?

Operational context, mission, use case

Clean & healthy teeth, gain of time, fashion bathroom

Functional
Electronic toothbrush

What ?

Service, function, task, operation, mode of operation

Brushing, speed regulating, brushing strength programming

Constructional

How ?

Component, device, configuration

Head, base, corpse, speed regulator

Elements of complex systems architecture – P. 34

Systems architecture frameworks Operational vision

Operational breakdown structure

E-toothbrushing contexts

Operational scenarios End user Shop

Types of operational contexts given by the needs

End user contexts

Maintenance contexts

Broken e-toothbrush

Repairing information

Health & care contexts

Fashion & glamour contexts

Efficiency & gain of time contexts

Problem detection

Repairing decision Repairing Repaired e-toothbrush

Electronic toothbrush
Efficient toothbrushing Toothbrushing analysis Quick morning brushing Quick brush filling

Operational contexts

Repairing

Operational vision: defines the intended objectives & uses and the ways of operating of the system relatively to the externally interfaced systems (customers, end users, etc.)
Elements of complex systems architecture – P. 35

Systems architecture frameworks Functional vision

Functional breakdown structure

Electronic toothbrushing

Tooth brushing

Brush brushing

Brushing programming

Internal technical functions

Types of functions classified by contexts

User functions

Toothbrushing reporting

Process organization

Electrical charging

Toothpaste filling

Technical functions

Service functions

Electronic toothbrush
Abstract functions
Brushing programming Toothbrushing reporting

• •

Brush brushing is a function Brush brushing using QTB123 brush is not a function That is to say not directly associated with a technical solution

Base waiting

Brush brushing

Functional vision: defines the abstract functions that are required to perform the missions of the system
Elements of complex systems architecture – P. 36

37 .Systems architecture frameworks Constructional vision Electronic toothbrush Product breakdown structure Toothbrush base Electronic toothbrush Toothbrush head Toothbrush base Toothbrush corpse Toothbrush head Toothbrush corpse Device organization Electrical alimentation Base shell Head corpse Brush Electronic toothbrush Mid-level or elementary components Head toothpaste channel • • • Head shell Brushing engine Hardware components Software components People Constructional vision: defines the real components that are going to perform the abstract functions Elements of complex systems architecture – P.

components & missions of a given system Elements of complex systems architecture – P. functions & components Functional breakdown structure Electronic toothbrushing Toothbrush base Toothbrush head Toothbrush corpse Toothbrushing Brushing programming Toothbrushing reporting Device organization Electronic toothbrush Operational breakdown structure E-toothbrushing contexts Toothbrushing Brushing programming Product breakdown structure Toothbrushing reporting Fashion & glamour contexts Health & care contexts Efficiency & gain of time contexts Toothbrush base Toothbrush corpse Toothbrush head Process organization Viewpoint traceability: tracing the coherent organization of functions. 38 .Systems architecture frameworks Relationships between systems visions Electronic toothbrush missions.

39 .Systems architecture frameworks Example of architectural framework (1) Classified according to their supervision frequency SAGACE framework What could be the problem? Action Control Control context context (scenario) (scenario) Tactics Management Management context context (scenario) (scenario) Strategy Strategic Strategic context context (scenario) (scenario) To expect Operational To do Functional What the system shall do? Function Function (dependency) (dependency) Task Task (operation) (operation) Operation Operation mode mode (scenario) (scenario) To be Constructional What the system shall be? Resource Resource (architecture) (architecture) Device Device (organization) (organization) Configuration Configuration (scenario) (scenario) Elements of complex systems architecture – P.

40 .Systems architecture frameworks Example of architectural framework (2) Classified according to the SysML expressivity SysML framework What could be the problem? States Life Life cycle cycle (behavior) (behavior) Statics Operational Operational contexts contexts (mission) (mission) Dynamics Operational Operational scenarios scenarios (scenario) (scenario) To expect Operational To do Functional What the system shall do? Operation Operation modes modes (behavior) (behavior) Functions Functions (operation) (operation) Function Function dynamics dynamics (scenario) (scenario) To be Constructional What the system shall be? Configurations Configurations (behavior) (behavior) Resources Resources (organization) (organization) Resource Resource dynamics dynamics (scenario) (scenario) Elements of complex systems architecture – P.

Systems architecture frameworks Example of architectural framework (3) Technical system Strategy Business objectives Human organization General management Responsability relationships Impacts on business processes Business processes Client relation management Responsability relationships Commercial management Operations Responsability relationships User interfaces Siebel Data Warehouse Other IT components Impacts on IT systems Information system management IT systems A classical enterprise architecture framework for information systems design Elements of complex systems architecture – P. 41 .

42 .Systems architecture frameworks Organization of a system model System level Operational structure Refinement Abstraction Allocation Functional structure Abstraction Allocation Constructional structure Abstraction Refinement Allocation Refinement Allocation Subsystem level Refinement Operational structure Abstraction Refinement Functional structure Abstraction Refinement Constructional structure Abstraction Module levels Refinement Operational structure Abstraction Allocation Functional structure Abstraction Allocation Constructional structure Abstraction Refinement Refinement Part level Operational structure Allocation Functional structure Allocation Constructional structure Two system modeling dimensions given respectively by the system hierarchy & the architectural framework Elements of complex systems architecture – P.

Contents 2nd topic 1st sub-topic 2nd sub-topic 3rd sub-topic Key systems architecture concepts What is a system? Systems architecture frameworks Systems architecture description P. 24 P. 43 . 33 P. 25 P. 43 Elements of complex systems architecture – P.

q. 44 .t) Special type of blocks Flow Partial function Elements of complex systems architecture – P.Systems architecture description What should be find in a system model? State machine • • • Integration operator: Internal Bloc Diagram Abstraction operators: association relationship Inter-vision relations: allocation relationship Port Transition + + Input (X) Internal (q) Algorithms: Sequence Diagram Output (Y) Y = F(X.

45 .Systems architecture description Organization of a system model Architectural Visions Requirements States Static structure Dynamical behavior Business data Operational vision Operational requirements (Requirements diagram) Operational contexts synthesis (State machine) Operational contexts (Bloc definition & internal bloc diagrams) Operational scenarios (Sequence diagram for each operational context) Operational data (Bloc definition diagram) Functional vision Functional requirements (Requirements diagram) Functional modes synthesis (State machine) Functional decomposition & interactions (Bloc definition & internal bloc diagrams) Functional behaviors (Sequence diagram for each function) Functional data (Bloc definition diagram) Constructional vision Constructional requirements (Requirements diagram) Configurations synthesis (State machine) Constructional decomposition & interactions (Bloc definition & internal bloc diagrams) Constructional behaviors (Sequence diagram for each resource) Constructional data (Bloc definition diagram) Typical structure of a SysML oriented system model at a given systemic level Elements of complex systems architecture – P.

46 .Systems architecture description Requirements diagram Requirements diagram Models requirements & their interdependencies Requirement Models a requirement Nesting Models a conjonctive decomposition relationship (A ∧ B → P) Derivation Models a logical consequence relationship (P → Q) Elements of complex systems architecture – P.

Systems architecture description Structural diagrams Bloc definition diagram Models a systemic organization (here a constructional decomposition) Bloc Models a systemic component Internal bloc diagram Models the internal relationships between different systemic components (here constructional relations) Flow port Models interaction points To describe static elements & relationships Elements of complex systems architecture – P. 47 Block instance Instances of blocs modeling specific systemic components .

48 Message Models an exchange of energy or information between two systemic components Bloc Models a systemic component .Systems architecture description Behavioral diagrams To express dynamic behaviors State machine Models the evolution of the states of a system Transition Models the change of state Sequence diagram Models the interactions between system components Operation Models an operational behavior State Models an operational context. a functional mode or a constructional configuration Elements of complex systems architecture – P.

Systems architecture description Business data & objects Bloc definition diagram Models business data & objects managed by the system Bloc Models a business object or data Association Models a relationship existing between two business data or objects Elements of complex systems architecture – P. 49 .

Systems architecture description Example of a system model (1) Electronic toothbrush The starting point: analyzing & defining the systemic perimeter of the system studied from an operational point of view Elements of complex systems architecture – P. 50 .

Systems architecture description Example of a system model (2) Electronic toothbrush States Static structure Dynamical behavior Operational vision Functional vision Constructional vision Elements of complex systems architecture – P. 51 .

Contents Introduction 1st topic 2nd topic 3rd topic Systems engineering context Key systems architecture concepts Some advanced topics Conclusion P. 71 Elements of complex systems architecture – P. 52 P. 52 . 24 P. 6 P. 3 P.

53 P. 53 . 52 P.Contents 3rd topic 1st sub-topic 2nd sub-topic Some advanced topics Design of families of systems Systems complexity measures P. 62 Elements of complex systems architecture – P.

54 . x = b = apple Weight (x) = m(x) g Visible light is just a part of the electro-magnetic spectrum Aristotle physics Teleonomy An apple belongs to earth and tries therefore always to get back to the earth Light belongs to sky and tries therefore always to get back to the sky Facts Gravity Cosmos Light Elements of complex systems architecture – P.b) 2 a = sun. b = planet planet sun Lorentz relativity Maxwell equations Galilean relativity a = earth.Design of families of systems Abstraction is a key lever for scientific progress … Abstraction Einstein relativity laws + Light speed c is independant of the Galilean observation referential Relativity speed(x) << c speed(x) ≈ c Modern physics Classical Galilean relativity & Newtonian mechanics Classical physics F(x) = m(x) γ(x) + Attraction(a.b) = G m(a)m(b) / d(a.

α (χ (a)) ⊂ a Gaps to fill Existing solution Abstraction Abstracting No properties should be lost α : C → A χ : C → A Two complementary ways for solving problems starting from known solutions Elements of complex systems architecture – P. 55 .Design of families of systems … and for system design Generalization C⊂A Existing solution Problems Gap analysis Properties may be lost Existing solution Model extension Gaps to reduce Galois correspondance Parametrization Existing solution ∀ c ∈ C. c ⊂ χ (α (c)) ∀ a ∈ A.

2004 Elements of complex systems architecture – P.Design of families of systems An example of families of systems 80% standard 20% spécifique Source: N. 56 . Lartigue. PSA Peugeot Citroën.

57 .b) Specific part Generic answer for all generic needs Solutions instanciation via specific development & parametrization of generic answer Customer needs 1 Customer needs 2 … Customer needs N Customer reality Generic_need(c.b) Abstraction allows to avoid product diversity which is the traditional problem of the traditional way of working (one context-optimized solution per customer) Elements of complex systems architecture – P.d) … … Generic_need(a.d) Generic_need(a.Design of families of systems What is behind: abstract architectures Abstract layer Specific Customer needs Abstraction of all customer needs Abstract generic solution Generic_Need(c.

Design of families of systems An optimization reformulation (1) Elements of complex systems architecture – P. 58 .

59 .Design of families of systems An optimization reformulation (2) Elements of complex systems architecture – P.

Design of families of systems Our hardware example revisited Unique Customized Options Flexible architecture How to organize a constructional architecture to implement the abstraction paradigm for families of systems in an hardware context already discussed Elements of complex systems architecture – P. 60 .

61 .Design of families of systems A software example Problem: manage the messages exchanges between N differents information systems CRM Logical link The CRM system is sending to the ERP & the PLM systems the « Starting date of a contract » ERP Logical link Message PLM ERP wrapper CRM wrapper PLM wrapper Relevant data EAI Relevant data Answer: a generic workflow management system A message of type A coming from the system X should be resent to the system Y with meta-data B Business rules Systems adresses Encapsulation of the relevant data with suitable meta-data Elements of complex systems architecture – P.

52 P. 53 P.Contents 3rd topic 1st sub-topic 2nd sub-topic Some advanced topics Design of families of systems Systems complexity measures P. 62 Elements of complex systems architecture – P. 62 .

63 .Systems complexity measures From recursive architectures to recursive graphs High level of a recursive architecture & graph Elements of complex systems architecture – P.

64 .Systems complexity measures From recursive architectures to recursive graphs Lower level of a recursive architecture & graph Elements of complex systems architecture – P.

the vertices of G of order i are obtained by partitionning the vertices of G of order i-1 when an arrow of G of order i is always an arrow of G of order i-1. G is a family of N usual graphs (Gi)i=1. G is just an ordinary graph.N = (Vi.Ai) such that one has for every i: • Vi is a partition of Vi-1. • when N > 1. • Ai ⊂ Ai-1 ..Systems complexity measures Recursive graphs Definition: a recursive graph G of order N is defined by setting: • when N = 1. • Hence. Passing from level i to level i-1 of in a recursive graph is called zooming within the graph Elements of complex systems architecture – P. 65 .

Systems complexity measures Weighted recursive graphs • Definition: A weighted recursive graph is defined by: • a recursive graph G = (Vi. πi is the vertex weight function on the i-th level Gi of G which is recursively defined by setting πi(V) = ∑ x∈V The weight of a « gluing » of vertices is the sum of the weights of the « glued » vertices … Elements of complex systems architecture – P. 66 πi (x) . • for i > 1.Ai)i=1.N. Weights are on the vertices ! . • a family (πi)i=1.N of vertex weight functions associated with each level of G which are defined by asking that: • π1 is any vertex weight function on the lower level G1 of G...

The complexity measure mn(G) of G of order is then the value defined by: mn(G) = ( ∑ π(x1) … π(xn) )1/n (x1. xn) path in G1 • Architectural complexity measure(s) of a system: Let S be a system defined from any architectural (that can typically be operational. functional or constructional) point of view. 67 . Elements of complex systems architecture – P. etc.). …. cost. The architectural complexity measure of order n of S is then the complexity measure of order n of the weighted recursive graph underlying to the considered architecture equiped with an initial weigthing of its primitive components which is proportional to what one wants to measure (effort.Systems complexity measures Architectural complexity measures • Complexity measure of a recursive graph: Let G be a weighted recursive graph whose lower (and first) level G1 = (V1.A1) is weighted by a weight function π.

The following properties of the complexity measures of a recursive graph weighted by a positive weight function π hold: • One always has mn(G)n ≤ mi(G)i mn-i(G)n-i for every i ≥ 1. there exist a value λ such that: mn(G) → λ n -> + ∝ λ can hence be interpreted as a kind of intrinsic complexity of G. Elements of complex systems architecture – P.. 2. The complexity measures mn(G) are linear. • By consequence.N) = mn((Gi)i=1. zoom independent and non destructive which means that one has: mn((Gi)i=1.N-1) when the passage from GN-1 to GN is realized by creating a complete graph with all the vertices of level N-1. 68 .Systems complexity measures Properties of our complexity measures 1..

X X X EAI X X X X X X Y X X X A n-th order spaghetti architecture (Kn) Elements of complex systems architecture – P. our complexity measures can be used in order to compare different architectural choices for chosing the « less » complex (see below for a typical architecture comparison example in an information systems context).Systems complexity measures Example of complexity measure use (1) In practice. 69 A n-th order EAI = (Enterprise Application Integration) architecture (Kn) .

e. the condition translates into: n2 X2 > 5n X2 + n XY. i. when n is big). 70 . to nX > 5X + Y or (n-5) X > Y or equivalently to the following relationship between X. Elements of complex systems architecture – P.e. it is a good choice to choose an EAI architecture (which seems reasonable). Y and n: Y/n < X Hence if the average complexity of the EAI bus per system connected to the bus is strictly less that the complexity of the systems.Systems complexity measures Example of complexity measure use (2) The limit complexity of the n-th order spaghetti architecture is: λ(Kn) = n X The limit complexity of the n-th order EAI architecture is: λ(En) = (X+Y)/2 + ( (X-Y)2 + 4 (n-1)X2)1/2 The EAI architecture is better than the spaghetti architecture iff one has: (n2 –5n +13/4)X2 + XY(5/2 –n) – 3/4Y2 > 0 Assymptotically (i.

71 Elements of complex systems architecture – P. 6 P. ? P.Contents Introduction 1st topic 2nd topic 3rd topic Systems engineering context Key systems architecture concepts Some advanced topics Conclusion P. 71 . 24 P. 3 P.

Word. SysML System verification Interface definition Simulink. 72 . Mathlab System integration System level Sub-system design Sub-systems level Sub-system construction Sub-system verification Time Key engineering challenge: a smooth design process at system level Elements of complex systems architecture – P.Conclusion What is the main system architecture challenge? (1) System validation Poor configuration management tools Tests Need analysis Requirement engineering tools System design Powerpoint.

measuring & dimensionning System verification Interface definition System integration System level One single « system » modeling & simuling tool Elements of complex systems architecture – P. 73 .Conclusion What is the main system architecture challenge? (2) • Key theoretical challenge: constructing an unified system theory based on an architectural perspective … • An unified formal behavioral system model • An unified formal point of view on architectural frameworks • … leading to unified system modeling tools at system level Need analysis Requirements referentials Configuration management tools System validation Formal verification & tests System design System model Simuling.

74 .End of the second day End of the course Elements of complex systems architecture – P.