Software development with components Component composition [Sommerville], chap 19.3 Composition Issues: Architectural mismatch [CrnkovicBook], chap 9 [GarlanArticle] Predictable composition

Software development with components Component composition [Sommerville], chap 19.3 Composition Issues: Architectural mismatch [CrnkovicBook], chap 9 [GarlanArticle] Predictable composition

Component composition The process of assembling components to create a system. Composition involves integrating components with each other and with the component infrastructure. The ways in which components are integrated with this infrastructure are specific for each component model Normally you also have to write ‘glue code’ to integrate components.

Types of composition Sequential composition where the composed components are executed in sequence. This involves composing the provides interfaces of each component. Glue code: calls component A, collects result, then calls component B with that result as parameter Hierarchical composition where one component calls on the services of another. The provides interface of one component is composed with the requires interface of another. Additive composition where the interfaces of two components are put together to create a new component.

Types of composition Figure 19.9 from [Sommerville]

Interface incompatibility When you desig components specially for composition, interfaces are designed to be compatible and composition is easy. When the components are developed independently for reuse, interfaces may be incompatible: Parameter incompatibility where operations have the same name but are of different types. Operation incompatibility where the names of operations in the composed interfaces are different. Operation incompleteness where the provides interface of one component is a subset of the requires interface of another Interface incompatibility is addressed by writing adaptors Adaptors address the problem of component incompatibility by reconciling the interfaces of the components that are composed. Different types of adaptor are required depending on the type of composition.

Example 1: Incompatible components Figure 19.10 from [Sommerville]

Example 1: Incompatible components addressFinder component: finds the address that matches a phone number mapper component: takes a post code and displays a street map of the area emergency operator: receives an emergency phone call, identifies the calling number and displays the street map of the address to send there an emergency vehicle Realised by composing an addressFinder with a mapper Problem: addressFinder returns a string describing the complete address, but the mapper needs only the post code An adaptor component called postCodeZipper takes the location data from address Finder and strips out the post code

Example 1: Composition through an adaptor The component postCodeStripper is the adaptor that facilitates the sequential composition of addressFinder and mapper components. address = addressFinder.location (phonenumber) ; postCode = postCodeStripper.getPostCode (address) ; mapper.displayMap(postCode, 10000)

Example 2: Adaptor for data collector There is an incompatibility between the provides interface of the sensor component with the requires interfaces of the data collection component Figure 19.11 from [Sommerville]

Interface semantics Component composition assumes you can tell from the component documentation whether the interfaces are compatible Syntactic compatibility: operation names and parameter types Semantic compatibility: the meaning of parameters is right You have to rely on component documentation to decide if interfaces that are syntactically compatible are actually compatible. Example: Consider an interface for a PhotoLibrary component:

Example: Photo library composition Compose a system that downloads images from a digital camera and stores them in a photograph library . The system user can provide additional info to catalog and describe the images Figure 19.12 from [Sommerville]

Example - formal description of photo library Figure 19.13 from [Sommerville]

Example - Photo library conditions explained As specified, the OCL associated with the Photo Library component states that: There must not be a photograph in the library with the same identifier as the photograph to be entered; The library must exist - assume that creating a library adds a single item to it; Each new entry increases the size of the library by 1; If you retrieve using the same identifier then you get back the photo that you added; If you look up the catalogue using that identifier, then you get back the catalogue entry that you made.

CBD Key points During the CBSE process, the processes of requirements engineering and system design are interleaved. Component composition is the process of ‘wiring’ components together to create a system. When composing reusable components, you normally have to write adaptors to reconcile different component interfaces. When choosing compositions, you have to consider required functionality, non-functional requirements and system evolution.

Software development with components Component composition [Sommerville], chap 19.3 Composition Issues: Architectural mismatch [CrnkovicBook], chap 9 [GarlanArticle] Predictable composition

Component Integration Integrating components can be illustrated as a mechanical process of “wiring” components together to form assemblies (connecting required and provided interfaces) Standardization in form of component models like EJB, CORBA and COM: reduces some of the integration problems

Component Composition Integrating (“wiring”) components is only a first step Composition (making components “play well together”) is still difficult: Problem goes beyond interface compatibility (syntactic and semantic) In many cases mismatches may be caused by low-level problems of interoperability, such as incompatibilities in programming languages, operating platforms, or database schemas. Architectural mismatch results from implicit and conflicting assumptions that designers of components make about the environment in which these components will operate

Architectural mismatch “Architectural mismatch stems from mismatched assumptions a reusable part makes about the structure of the system it is to be part of. These assumptions often conflict with the assumptions of other parts and are almost always implicit, making them extremely difficult to analyze before building the system.” D. Garlan, R. Allen and J. Ockerbloom. “Architectural Mismatch: Why Reuse is So Hard,” IEEE Software, 12(6):17-26, November 1995 http://www.cs.cmu.edu/afs/cs/project/able/ftp/archmismatch-icse17/archmismatch-icse17.pdf

Component composition issues “Architectural mismatch stems from mismatched assumptions a reusable part makes about the structure of the system it is to be part of” four classes of structural assumptions The nature of components (infrastructure, control model, and data model) The nature of connectors (protocols and data models) The architecture of the assemblies (constraints on interactions) The run-time construction process (order of instantiations).

Architectural mismatch examples 2 Cases D. Garlan, R. Allen and J. Ockerbloom “Architectural Mismatch: Why Reuse is So Hard” AESOP P. Inverardi, A.L. Wolf, and D. Yankelevich, Static Checking of System Behaviors Using Derived Component Assumptions Compressing proxy

AESOP Example AESOP: developed by CMU, an environment that generates development environments tailored for systems of a particular style Development approach: integrate existing components Components: A object oriented database (OBST – public domain system) GUI toolkit (Interviews and Unidraw – Stanford university) An event-based tool-integration mechanism (SoftBench from HP) A RPC mechanism (Mach RPC Interface Generator – CMU) Advantages: Stable tools, used in several other projects All tools implemented in C/C++ AND with source code available

Problems in AESOP integration (1) Assumptions about the components: infrastructure: Components assume they have a certain infrastructure, but it is not available Ex: SoftBench assumes all components have their own GUI interfaces and hence have the X-library’s communication primitives loaded => all components had to load X library even if they do not need it otherwise => code bloat Components assume that they should provide a certain infrastructure, which the application does not need Ex: OBST provides many functions, Aesop needs only few of these

Problems in AESOP integration (2) Assumptions about the components: control model: Different packages make different assumptions about which part holds the main thread of control Ex: SoftBench, InterViews and MIG all use event loops that are incompatible with each other => rewrite InterViews event loop

Problems in AESOP integration (3) Assumptions about the components: data model: Different components assume different things about the nature of data Ex: Unidraw maintain a hierarchichal model for its objects but allows only top-level objects to be manipulated by users, Aesop requires that both parent and child objects can be manipulated => rewrite Unidraw hierarchy

Problems in AESOP integration (4) Assumptions about the connectors: protocols SoftBench handles RPC by mapping it within the event framework through 2 events (the request and the reply) => the caller becomes more complicated => use Mach RPC instead of Softbench

Problems in AESOP integration (5) Assumptions about the connectors: data model Components have different assumptions on what comes over a connection Ex: SoftBench: Strings; MIG: C data; OBST C++ data. translation between formats needed => performance bottleneck

Problems in AESOP integration (6) Assumptions about the global architecture OBST assumes that all communications occurs in a star configuration, with itself in the center, Assumes independence of client tools and provides a transaction protocol per single tool, not per combination of tools Aesop’s communication structure is a more general graph (tools cooperate also directly) => OBST’s transaction mechanism can result in deadlock or database inconsistency => write own transaction mechanism Assumes independence of client tools and provides a transaction protocol per single tool, not per combination of tools

Problems in AESOP integration (7) Assumptions about the building process Assumptions about the library infrastructure Assumptions about a generic language (C++) Assumptions about a tool specific language Some component A may have other expectations on the generated code of another component B as B itself Assumptions about the library infrastructure Assumptions about a generic language (C++) Assumptions about a tool specific language ► Combination is fatal: ■ Some component A may have other expectations on the generated code of another component B as B itself ■ Then, the developer has to patch the generated code of A with patch scripts (another translation component)

Consequences for Aesop Effort: Estimated: one person-year Reality: 5 person-years Code bloat Poor performance: due to overhead of the tool-to-database communication and excessive code Need to modify existing packages: ex. Event loop of SoftBench and Interview Need to reimplement some existing functions: OBST transaction mechanism, Hierarchical nested view management in InterView

Component integration issues “Architectural mismatch stems from mismatched assumptions a reusable part makes about the structure of the system it is to be part of” four classes of structural assumptions The nature of components (infrastructure, control model, and data model) The nature of connectors (protocols and data models) The architecture of the assemblies (constraints on interactions) The run-time construction process (order of instantiations).

Recommended practice To avoid architectural mismatch: Make assumptions explicit Use orthogonal loosely-coupled components Use bridging techniques to solve mismatches Provide design and composition guidelines

Compressing proxy example Problem: adding data compression to a web server, in order to improve performance Characteristics of web server: Standard CERN HTTP server Pipes-and-filters architecture Proposed solution: Add an external filter to compress data using gzip External compressing filter communicates with the web server through Unix pipes (read and write data) A pseudo filter (adaptor) is added to work with the external compressing filter

Compressing proxy example Filter Pseudo Filter (Adaptor) gzip 1 2 3 4 Compressing Proxy Process Component Channel Function call interface UNIX pipe interface 1 contains a graphical representation of the Compressing Proxy. Data to be transmitted over the Web enters the proxy by way of the filter on the left. It is then fed into another filter that acts as an adaptor which supports the use of the gzip application, which performs data compression. Careful analysis revealed the potential for deadlock because the adaptor blocked while supplying data to gzip. The adaptor could not then receive any compressed data until all the data had been read by gzip. Gzip was not allowed to offload zipped data over connection 3 until all the data had been read from the adaptor via connection 2. This meant that any attempt to process a file larger than the capacity of the gzip buffer would cause the system to deadlock. Based on this finding the adaptor was replaced by a non-blocking version and the proxy functioned as intended. Figure 9.1 from [CrnkovicBook]

Compressing proxy issues HTTP server filters are forced to read when data are pushed at them Unix filters choose when to read data -> gzip may block Normal scenario: Adaptor passes the data from input filter to gzip, and when the stream is closed, reads the data from gzip and writes in the output filter Problematic scenario: If the input file is large, because gzip has a limited internal buffer, it may attempt to write a portion of compressed data before the adaptor is ready: => deadlock (gzip blocks and adapter is blocked) Solution: Adaptor should handle data incrementally and use unblocking read and write

Lessons learned Formal architectural description and analysis to uncover what they call “behavioral mismatch” Not a component mismatch (components can be “wired” together) Components must express behavioural properties: assumptions made about it’s environment such as data formats or buffer sizes Its effects on the environment The success of a component marketplace depends on: Having trustworthy claims for component properties Having global analysis techniques to support reasoning about the emergent properties of assemblies

From Integration to Composition All assemblies are potential subsystem Predicting the emergent behavior of assemblies The result of component composition is a component assembly which can be used as a part of a larger composition Composition goes beyond integration by allowing prediction of the emergent behavior of assemblies Compositional reasoning: if we know the properties of components c1 and c2, then we can define a reasoning function f such that f(c1,c2) yields a property of an assembly comprising c1 and c2 Component composition goes one step further than integration in that the resulting composition can be a part of larger assemblies. Composition is based on the ability to assign properties to the whole based on the properties of the parts. composition is distinguished from integration primarily by the fact that composition also focuses on emergent assembly-level behaviour, making certain that the assembly will perform as desired and that it could be used as a building block in a larger system. The constituent components must not only plug together, they must play well together. In analogy, consider the incompatibility of connecting a very powerful audio amplifier to low wattage speakers. The speakers will plug in with no problem, and at low volumes will probably function acceptably but if the volume is raised the speakers will most likely be destroyed.

Predictable Assembly from Certifiable Components What types of system quality attributes are developers interested in predicting? What types of analysis techniques support reasoning about these quality attributes, and what component property values do they require as input parameters? How are these component properties specified, measured, and certified? These three questions are interdependent. The types of compositional reasoning which can be accomplished ultimately depend upon the types of component properties which can be measured. Conversely, it is the reasoning techniques which determine what component properties are material in the first place. Therefore, the answers to these three questions, which are mutually informing (and constraining), will provide a foundation for a sustainable improvement in predicting the properties of component assemblies and for confidence in the software components which make up these assemblies. However, answering these questions will be an ongoing process: new prediction models will require new and/or improved component measures, which will in turn lead to more accurate prediction, and to a demand for better or additional prediction models, more precise component measures, and so forth.

Summary Component-based development includes: Component development (design for reuse) System development (design with reuse) System composition/integration may show problems: Interface incompatibility Architectural/ mismatch Emergent properties of component assembly: Certified component Analysis techniques - compositional reasoning