Data modeling

From Wikipedia, the free encyclopedia

 

The data modeling process. The figure illustrates the way data models are developed and used today . A conceptual data model is developed based on the data requirements for the application that is being developed, perhaps in the context of an activity model. The data model will normally consist of entity types, attributes, relationships, integrity rules, and the definitions of those objects. This is then used as the start point for interface or database design.

 

Data modeling in software engineering is the process of creating a data model for an information system by applying certain formal techniques.

Overview

Data modeling is a process used to define and analyze data requirements needed to support the business processes within the scope of corresponding information systems in organizations. Therefore, the process of data modeling involves professional data modelers working closely with business stakeholders, as well as potential users of the information system. 

There are three different types of data models produced while progressing from requirements to the actual database to be used for the information system. The data requirements are initially recorded as a conceptual data model which is essentially a set of technology independent specifications about the data and is used to discuss initial requirements with the business stakeholders. The conceptual model is then translated into a logical data model, which documents structures of the data that can be implemented in databases. Implementation of one conceptual data model may require multiple logical data models. The last step in data modeling is transforming the logical data model to a physical data model that organizes the data into tables, and accounts for access, performance and storage details. Data modeling defines not just data elements, but also their structures and the relationships between them.

Data modeling techniques and methodologies are used to model data in a standard, consistent, predictable manner in order to manage it as a resource. The use of data modeling standards is strongly recommended for all projects requiring a standard means of defining and analyzing data within an organization, e.g., using data modeling:

• to assist business analysts, programmers, testers, manual writers, IT package selectors, engineers, managers, related organizations and clients to understand and use an agreed semi-formal model the concepts of the organization and how they relate to one another
• to manage data as a resource
• for the integration of information systems
• for designing databases/data warehouses (aka data repositories)

Data modeling may be performed during various types of projects and in multiple phases of projects. Data models are progressive; there is no such thing as the final data model for a business or application. Instead a data model should be considered a living document that will change in response to a changing business. The data models should ideally be stored in a repository so that they can be retrieved, expanded, and edited over time. Whitten et al. (2004) determined two types of data modeling:

• Strategic data modeling: This is part of the creation of an information systems strategy, which defines an overall vision and architecture for information systems. Information technology engineering is a methodology that embraces this approach.
• Data modeling during systems analysis: In systems analysis logical data models are created as part of the development of new databases.

Data modeling is also used as a technique for detailing business requirements for specific databases. It is sometimes called database modeling because a data model is eventually implemented in a database.

Data modeling topics

Data models

How data models deliver benefit.

 

Data models provide a framework for data to be used within information systems by providing specific definition and format. If a data model is used consistently across systems then compatibility of data can be achieved. If the same data structures are used to store and access data then different applications can share data seamlessly. The results of this are indicated in the diagram. However, systems and interfaces are often expensive to build, operate, and maintain. They may also constrain the business rather than support it. This may occur when the quality of the data models implemented in systems and interfaces is poor.

Some common problems found in data models are:

• Business rules, specific to how things are done in a particular place, are often fixed in the structure of a data model. This means that small changes in the way business is conducted lead to large changes in computer systems and interfaces. So, business rules need to be implemented in a flexible way that does not result in complicated dependencies, rather the data model should be flexible enough so that changes in the business can be implemented within the data model in a relatively quick and efficient way.
• Entity types are often not identified, or are identified incorrectly. This can lead to replication of data, data structure and functionality, together with the attendant costs of that duplication in development and maintenance.Therefore, data definitions should be made as explicit and easy to understand as possible to minimize misinterpretation and duplication.
• Data models for different systems are arbitrarily different. The result of this is that complex interfaces are required between systems that share data. These interfaces can account for between 25-70% of the cost of current systems. Required interfaces should be considered inherently while designing a data model, as a data model on its own would not be usable without interfaces within different systems.
• Data cannot be shared electronically with customers and suppliers, because the structure and meaning of data has not been standardised. To obtain optimal value from an implemented data model, it is very important to define standards that will ensure that data models will both meet business needs and be consistent.

Conceptual, logical and physical schemas

The ANSI/SPARC three level architecture. This shows that a data model can be an external model (or view), a conceptual model, or a physical model. This is not the only way to look at data models, but it is a useful way, particularly when comparing models.

 

In 1975 ANSI described three kinds of data-model instance:

• Conceptual schema: describes the semantics of a domain (the scope of the model). For example, it may be a model of the interest area of an organization or of an industry. This consists of entity classes, representing kinds of things of significance in the domain, and relationships assertions about associations between pairs of entity classes. A conceptual schema specifies the kinds of facts or propositions that can be expressed using the model. In that sense, it defines the allowed expressions in an artificial "language" with a scope that is limited by the scope of the model. Simply described, a conceptual schema is the first step in organizing the data requirements.
• Logical schema: describes the structure of some domain of information. This consists of descriptions of (for example) tables, columns, object-oriented classes, and XML tags. The logical schema and conceptual schema are sometimes implemented as one and the same.
• Physical schema: describes the physical means used to store data. This is concerned with partitions, CPUs, tablespaces, and the like.

According to ANSI, this approach allows the three perspectives to be relatively independent of each other. Storage technology can change without affecting either the logical or the conceptual schema. The table/column structure can change without (necessarily) affecting the conceptual schema. In each case, of course, the structures must remain consistent across all schemas of the same data model.

Data modeling process

Data modeling in the context of Business Process Integration.

 

In the context of business process integration (see figure), data modeling complements business process modeling, and ultimately results in database generation.

The process of designing a database involves producing the previously described three types of schemas - conceptual, logical, and physical. The database design documented in these schemas are converted through a Data Definition Language, which can then be used to generate a database. A fully attributed data model contains detailed attributes (descriptions) for every entity within it. The term "database design" can describe many different parts of the design of an overall database system. Principally, and most correctly, it can be thought of as the logical design of the base data structures used to store the data. In the relational model these are the tables and views. In an object database the entities and relationships map directly to object classes and named relationships. However, the term "database design" could also be used to apply to the overall process of designing, not just the base data structures, but also the forms and queries used as part of the overall database application within the Database Management System or DBMS.

In the process, system interfaces account for 25% to 70% of the development and support costs of current systems. The primary reason for this cost is that these systems do not share a common data model. If data models are developed on a system by system basis, then not only is the same analysis repeated in overlapping areas, but further analysis must be performed to create the interfaces between them. Most systems within an organization contain the same basic data, redeveloped for a specific purpose. Therefore, an efficiently designed basic data model can minimize rework with minimal modifications for the purposes of different systems within the organization

Modeling methodologies

Data models represent information areas of interest. While there are many ways to create data models, according to Len Silverston (1997) only two modeling methodologies stand out, top-down and bottom-up:

• Bottom-up models or View Integration models are often the result of a reengineering effort. They usually start with existing data structures forms, fields on application screens, or reports. These models are usually physical, application-specific, and incomplete from an enterprise perspective. They may not promote data sharing, especially if they are built without reference to other parts of the organization.
• Top-down logical data models, on the other hand, are created in an abstract way by getting information from people who know the subject area. A system may not implement all the entities in a logical model, but the model serves as a reference point or template.

Sometimes models are created in a mixture of the two methods: by considering the data needs and structure of an application and by consistently referencing a subject-area model. Unfortunately, in many environments the distinction between a logical data model and a physical data model is blurred. In addition, some CASE tools don't make a distinction between logical and physical data models.

Entity relationship diagrams

Example of an IDEF1X Entity relationship diagrams used to model IDEF1X itself. The name of the view is mm. The domain hierarchy and constraints are also given. The constraints are expressed as sentences in the formal theory of the meta model.

 

There are several notations for data modeling. The actual model is frequently called "Entity relationship model", because it depicts data in terms of the entities and relationships described in the data. An entity-relationship model (ERM) is an abstract conceptual representation of structured data. Entity-relationship modeling is a relational schema database modeling method, used in software engineering to produce a type of conceptual data model (or semantic data model) of a system, often a relational database, and its requirements in a top-down fashion.

These models are being used in the first stage of information system design during the requirements analysis to describe information needs or the type of information that is to be stored in a database. The data modeling technique can be used to describe any ontology (i.e. an overview and classifications of used terms and their relationships) for a certain universe of discourse i.e. area of interest.

Several techniques have been developed for the design of data models. While these methodologies guide data modelers in their work, two different people using the same methodology will often come up with very different results. Most notable are:

• Bachman diagrams
• Barker's notation
• Chen's Notation
• Data Vault Modeling
• Extended Backus–Naur form
• IDEF1X
• Object-relational mapping
• Object-Role Modeling
• Relational Model
• Relational Model/Tasmania

Generic data modeling

Example of a Generic data model.

 

Generic data models are generalizations of conventional data models. They define standardized general relation types, together with the kinds of things that may be related by such a relation type. The definition of generic data model is similar to the definition of a natural language. For example, a generic data model may define relation types such as a 'classification relation', being a binary relation between an individual thing and a kind of thing (a class) and a 'part-whole relation', being a binary relation between two things, one with the role of part, the other with the role of whole, regardless the kind of things that are related.

Given an extensible list of classes, this allows the classification of any individual thing and to specify part-whole relations for any individual object. By standardization of an extensible list of relation types, a generic data model enables the expression of an unlimited number of kinds of facts and will approach the capabilities of natural languages. Conventional data models, on the other hand, have a fixed and limited domain scope, because the instantiation (usage) of such a model only allows expressions of kinds of facts that are predefined in the model.

Semantic data modeling

The logical data structure of a DBMS, whether hierarchical, network, or relational, cannot totally satisfy the requirements for a conceptual definition of data because it is limited in scope and biased toward the implementation strategy employed by the DBMS. That is unless the semantic data model is implemented in the database on purpose, a choice which may slightly impact performance but generally vastly improves productivity. 

Semantic data models.

 

Therefore, the need to define data from a conceptual view has led to the development of semantic data modeling techniques. That is, techniques to define the meaning of data within the context of its interrelationships with other data. As illustrated in the figure the real world, in terms of resources, ideas, events, etc., are symbolically defined within physical data stores. A semantic data model is an abstraction which defines how the stored symbols relate to the real world. Thus, the model must be a true representation of the real world.

A semantic data model can be used to serve many purposes, such as:

• planning of data resources
• building of shareable databases
• evaluation of vendor software
• integration of existing databases

The overall goal of semantic data models is to capture more meaning of data by integrating relational concepts with more powerful abstraction concepts known from the Artificial Intelligence field. The idea is to provide high level modeling primitives as integral part of a data model in order to facilitate the representation of real world situations.

Interpreter (computing)

From Wikipedia, the free encyclopedia

In computer science, an interpreter is a computer program that directly executes instructions written in a programming or scripting language, without requiring them previously to have been compiled into a machine language program. An interpreter generally uses one of the following strategies for program execution:

1. Parse the source code and perform its behavior directly;
2. Translate source code into some efficient intermediate representation and immediately execute this;
3. Explicitly execute stored precompiled code made by a compiler which is part of the interpreter system.

Early versions of Lisp programming language and Dartmouth BASIC would be examples of the first type. Perl, Python, MATLAB, and Ruby are examples of the second, while UCSD Pascal is an example of the third type. Source programs are compiled ahead of time and stored as machine independent code, which is then linked at run-time and executed by an interpreter and/or compiler (for JIT systems). Some systems, such as Smalltalk and contemporary versions of BASIC and Java may also combine two and three. Interpreters of various types have also been constructed for many languages traditionally associated with compilation, such as Algol, Fortran, Cobol and C/C++.

While interpretation and compilation are the two main means by which programming languages are implemented, they are not mutually exclusive, as most interpreting systems also perform some translation work, just like compilers. The terms "interpreted language" or "compiled language" signify that the canonical implementation of that language is an interpreter or a compiler, respectively. A high level language is ideally an abstraction independent of particular implementations.

History

Interpreters were used as early as 1952 to ease programming within the limitations of computers at the time (e.g. a shortage of program storage space, or no native support for floating point numbers). Interpreters were also used to translate between low-level machine languages, allowing code to be written for machines that were still under construction and tested on computers that already existed. The first interpreted high-level language was Lisp. Lisp was first implemented in 1958 by Steve Russell on an IBM 704 computer. Russell had read John McCarthy's paper, and realized (to McCarthy's surprise) that the Lisp eval function could be implemented in machine code. The result was a working Lisp interpreter which could be used to run Lisp programs, or more properly, "evaluate Lisp expressions".

Compilers versus interpreters

An illustration of the linking process. Object files and static libraries are assembled into a new library or executable

 

Programs written in a high level language are either directly executed by some kind of interpreter or converted into machine code by a compiler (and assembler and linker) for the CPU to execute.

While compilers (and assemblers) generally produce machine code directly executable by computer hardware, they can often (optionally) produce an intermediate form called object code. This is basically the same machine specific code but augmented with a symbol table with names and tags to make executable blocks (or modules) identifiable and relocatable. Compiled programs will typically use building blocks (functions) kept in a library of such object code modules. A linker is used to combine (pre-made) library files with the object file(s) of the application to form a single executable file. The object files that are used to generate an executable file are thus often produced at different times, and sometimes even by different languages (capable of generating the same object format).

A simple interpreter written in a low level language (e.g. assembly) may have similar machine code blocks implementing functions of the high level language stored, and executed when a function's entry in a look up table points to that code. However, an interpreter written in a high level language typically uses another approach, such as generating and then walking a parse tree, or by generating and executing intermediate software-defined instructions, or both.

Thus, both compilers and interpreters generally turn source code (text files) into tokens, both may (or may not) generate a parse tree, and both may generate immediate instructions (for a stack machine, quadruple code, or by other means). The basic difference is that a compiler system, including a (built in or separate) linker, generates a stand-alone machine code program, while an interpreter system instead performs the actions described by the high level program.

A compiler can thus make almost all the conversions from source code semantics to the machine level once and for all (i.e. until the program has to be changed) while an interpreter has to do some of this conversion work every time a statement or function is executed. However, in an efficient interpreter, much of the translation work (including analysis of types, and similar) is factored out and done only the first time a program, module, function, or even statement, is run, thus quite akin to how a compiler works. However, a compiled program still runs much faster, under most circumstances, in part because compilers are designed to optimize code, and may be given ample time for this. This is especially true for simpler high level languages without (many) dynamic data structures, checks, or type-checks.

In traditional compilation, the executable output of the linkers (.exe files or .dll files or a library, see picture) is typically relocatable when run under a general operating system, much like the object code modules are but with the difference that this relocation is done dynamically at run time, i.e. when the program is loaded for execution. On the other hand, compiled and linked programs for small embedded systems are typically statically allocated, often hard coded in a NOR flash memory, as there is often no secondary storage and no operating system in this sense. 

Historically, most interpreter-systems have had a self-contained editor built in. This is becoming more common also for compilers (then often called an IDE), although some programmers prefer to use an editor of their choice and run the compiler, linker and other tools manually. Historically, compilers predate interpreters because hardware at that time could not support both the interpreter and interpreted code and the typical batch environment of the time limited the advantages of interpretation.

Development cycle

During the software development cycle, programmers make frequent changes to source code. When using a compiler, each time a change is made to the source code, they must wait for the compiler to translate the altered source files and link all of the binary code files together before the program can be executed. The larger the program, the longer the wait. By contrast, a programmer using an interpreter does a lot less waiting, as the interpreter usually just needs to translate the code being worked on to an intermediate representation (or not translate it at all), thus requiring much less time before the changes can be tested. Effects are evident upon saving the source code and reloading the program. Compiled code is generally less readily debugged as editing, compiling, and linking are sequential processes that have to be conducted in the proper sequence with a proper set of commands. For this reason, many compilers also have an executive aid, known as a Make file and program. The Make file lists compiler and linker command lines and program source code files, but might take a simple command line menu input (e.g. "Make 3") which selects the third group (set) of instructions then issues the commands to the compiler, and linker feeding the specified source code files.

Distribution

A compiler converts source code into binary instruction for a specific processor's architecture, thus making it less portable. This conversion is made just once, on the developer's environment, and after that the same binary can be distributed to the user's machines where it can be executed without further translation. A cross compiler can generate binary code for the user machine even if it has a different processor than the machine where the code is compiled. 

An interpreted program can be distributed as source code. It needs to be translated in each final machine, which takes more time but makes the program distribution independent of the machine's architecture. However, the portability of interpreted source code is dependent on the target machine actually having a suitable interpreter. If the interpreter needs to be supplied along with the source, the overall installation process is more complex than delivery of a monolithic executable since the interpreter itself is part of what need be installed. 

The fact that interpreted code can easily be read and copied by humans can be of concern from the point of view of copyright. However, various systems of encryption and obfuscation exist. Delivery of intermediate code, such as bytecode, has a similar effect to obfuscation, but bytecode could be decoded with a decompiler or disassembler.

Efficiency

The main disadvantage of interpreters is that an interpreted program typically runs slower than if it had been compiled. The difference in speeds could be tiny or great; often an order of magnitude and sometimes more. It generally takes longer to run a program under an interpreter than to run the compiled code but it can take less time to interpret it than the total time required to compile and run it. This is especially important when prototyping and testing code when an edit-interpret-debug cycle can often be much shorter than an edit-compile-run-debug cycle.

Interpreting code is slower than running the compiled code because the interpreter must analyze each statement in the program each time it is executed and then perform the desired action, whereas the compiled code just performs the action within a fixed context determined by the compilation. This run-time analysis is known as "interpretive overhead". Access to variables is also slower in an interpreter because the mapping of identifiers to storage locations must be done repeatedly at run-time rather than at compile time.

There are various compromises between the development speed when using an interpreter and the execution speed when using a compiler. Some systems (such as some Lisps) allow interpreted and compiled code to call each other and to share variables. This means that once a routine has been tested and debugged under the interpreter it can be compiled and thus benefit from faster execution while other routines are being developed. Many interpreters do not execute the source code as it stands but convert it into some more compact internal form. Many BASIC interpreters replace keywords with single byte tokens which can be used to find the instruction in a jump table. A few interpreters, such as the PBASIC interpreter, achieve even higher levels of program compaction by using a bit-oriented rather than a byte-oriented program memory structure, where commands tokens occupy perhaps 5 bits, nominally "16-bit" constants are stored in a variable-length code requiring 3, 6, 10, or 18 bits, and address operands include a "bit offset". Many BASIC interpreters can store and read back their own tokenized internal representation.

Toy expression interpreter

An interpreter might well use the same lexical analyzer and parser as the compiler and then interpret the resulting abstract syntax tree. Example data type definitions for the latter, and a toy interpreter for syntax trees obtained from C expressions are shown in the box.

Regression

Interpretation cannot be used as the sole method of execution: even though an interpreter can itself be interpreted and so on, a directly executed program is needed somewhere at the bottom of the stack because the code being interpreted is not, by definition, the same as the machine code that the CPU can execute.

Variations

Bytecode interpreters

There is a spectrum of possibilities between interpreting and compiling, depending on the amount of analysis performed before the program is executed. For example, Emacs Lisp is compiled to bytecode, which is a highly compressed and optimized representation of the Lisp source, but is not machine code (and therefore not tied to any particular hardware). This "compiled" code is then interpreted by a bytecode interpreter (itself written in C). The compiled code in this case is machine code for a virtual machine, which is implemented not in hardware, but in the bytecode interpreter. Such compiling interpreters are sometimes also called compreters. In a bytecode interpreter each instruction starts with a byte, and therefore bytecode interpreters have up to 256 instructions, although not all may be used. Some bytecodes may take multiple bytes, and may be arbitrarily complicated. 

Control tables - that do not necessarily ever need to pass through a compiling phase - dictate appropriate algorithmic control flow via customized interpreters in similar fashion to bytecode interpreters.

Threaded code interpreters

Threaded code interpreters are similar to bytecode interpreters but instead of bytes they use pointers. Each "instruction" is a word that points to a function or an instruction sequence, possibly followed by a parameter. The threaded code interpreter either loops fetching instructions and calling the functions they point to, or fetches the first instruction and jumps to it, and every instruction sequence ends with a fetch and jump to the next instruction. Unlike bytecode there is no effective limit on the number of different instructions other than available memory and address space. The classic example of threaded code is the Forth code used in Open Firmware systems: the source language is compiled into "F code" (a bytecode), which is then interpreted by a virtual machine.

Abstract syntax tree interpreters

In the spectrum between interpreting and compiling, another approach is to transform the source code into an optimized abstract syntax tree (AST), then execute the program following this tree structure, or use it to generate native code just-in-time. In this approach, each sentence needs to be parsed just once. As an advantage over bytecode, the AST keeps the global program structure and relations between statements (which is lost in a bytecode representation), and when compressed provides a more compact representation. Thus, using AST has been proposed as a better intermediate format for just-in-time compilers than bytecode. Also, it allows the system to perform better analysis during runtime. 

However, for interpreters, an AST causes more overhead than a bytecode interpreter, because of nodes related to syntax performing no useful work, of a less sequential representation (requiring traversal of more pointers) and of overhead visiting the tree.

Just-in-time compilation

Further blurring the distinction between interpreters, bytecode interpreters and compilation is just-in-time compilation (JIT), a technique in which the intermediate representation is compiled to native machine code at runtime. This confers the efficiency of running native code, at the cost of startup time and increased memory use when the bytecode or AST is first compiled. Adaptive optimization is a complementary technique in which the interpreter profiles the running program and compiles its most frequently executed parts into native code. Both techniques are a few decades old, appearing in languages such as Smalltalk in the 1980s.

Just-in-time compilation has gained mainstream attention amongst language implementers in recent years, with Java, the .NET Framework, most modern JavaScript implementations, and Matlab now including JITs.

Self-interpreter

A self-interpreter is a programming language interpreter written in a programming language which can interpret itself; an example is a BASIC interpreter written in BASIC. Self-interpreters are related to self-hosting compilers.

If no compiler exists for the language to be interpreted, creating a self-interpreter requires the implementation of the language in a host language (which may be another programming language or assembler). By having a first interpreter such as this, the system is bootstrapped and new versions of the interpreter can be developed in the language itself. It was in this way that Donald Knuth developed the TANGLE interpreter for the language WEB of the industrial standard TeX typesetting system.

Defining a computer language is usually done in relation to an abstract machine (so-called operational semantics) or as a mathematical function (denotational semantics). A language may also be defined by an interpreter in which the semantics of the host language is given. The definition of a language by a self-interpreter is not well-founded (it cannot define a language), but a self-interpreter tells a reader about the expressiveness and elegance of a language. It also enables the interpreter to interpret its source code, the first step towards reflective interpreting.

An important design dimension in the implementation of a self-interpreter is whether a feature of the interpreted language is implemented with the same feature in the interpreter's host language. An example is whether a closure in a Lisp-like language is implemented using closures in the interpreter language or implemented "manually" with a data structure explicitly storing the environment. The more features implemented by the same feature in the host language, the less control the programmer of the interpreter has; a different behavior for dealing with number overflows cannot be realized if the arithmetic operations are delegated to corresponding operations in the host language.

Some languages have an elegant self-interpreter, such as Lisp or Prolog. Much research on self-interpreters (particularly reflective interpreters) has been conducted in the Scheme programming language, a dialect of Lisp. In general, however, any Turing-complete language allows writing of its own interpreter. Lisp is such a language, because Lisp programs are lists of symbols and other lists. XSLT is such a language, because XSLT programs are written in XML. A sub-domain of meta-programming is the writing of domain-specific languages (DSLs).

Clive Gifford introduced a measure quality of self-interpreter (the eigenratio), the limit of the ratio between computer time spent running a stack of N self-interpreters and time spent to run a stack of N − 1 self-interpreters as N goes to infinity. This value does not depend on the program being run.

The book Structure and Interpretation of Computer Programs presents examples of meta-circular interpretation for Scheme and its dialects. Other examples of languages with a self-interpreter are Forth and Pascal.

Microcode

Microcode is a very commonly used technique "that imposes an interpreter between the hardware and the architectural level of a computer". As such, the microcode is a layer of hardware-level instructions that implement higher-level machine code instructions or internal state machine sequencing in many digital processing elements. Microcode is used in general-purpose central processing units, as well as in more specialized processors such as microcontrollers, digital signal processors, channel controllers, disk controllers, network interface controllers, network processors, graphics processing units, and in other hardware.

Microcode typically resides in special high-speed memory and translates machine instructions, state machine data or other input into sequences of detailed circuit-level operations. It separates the machine instructions from the underlying electronics so that instructions can be designed and altered more freely. It also facilitates the building of complex multi-step instructions, while reducing the complexity of computer circuits. Writing microcode is often called microprogramming and the microcode in a particular processor implementation is sometimes called a microprogram.

More extensive microcoding allows small and simple microarchitectures to emulate more powerful architectures with wider word length, more execution units and so on, which is a relatively simple way to achieve software compatibility between different products in a processor family.

Applications

• Interpreters are frequently used to execute command languages, and glue languages since each operator executed in command language is usually an invocation of a complex routine such as an editor or compiler.
• Self-modifying code can easily be implemented in an interpreted language. This relates to the origins of interpretation in Lisp and artificial intelligence research.
• Virtualization. Machine code intended for a hardware architecture can be run using a virtual machine. This is often used when the intended architecture is unavailable, or among other uses, for running multiple copies.
• Sandboxing: While some types of sandboxes rely on operating system protections, an interpreter or virtual machine is often used. The actual hardware architecture and the originally intended hardware architecture may or may not be the same. This may seem pointless, except that sandboxes are not compelled to actually execute all the instructions the source code it is processing. In particular, it can refuse to execute code that violates any security constraints it is operating under.
• Emulators for running computer software written for obsolete and unavailable hardware on more modern equipment.

Virtual machine

From Wikipedia, the free encyclopedia

 

In computing, a virtual machine (VM) is an emulation of a computer system. Virtual machines are based on computer architectures and provide functionality of a physical computer. Their implementations may involve specialized hardware, software, or a combination.

There are different kinds of virtual machines, each with different functions:

• System virtual machines (also termed full virtualization VMs) provide a substitute for a real machine. They provide functionality needed to execute entire operating systems. A hypervisor uses native execution to share and manage hardware, allowing for multiple environments which are isolated from one another, yet exist on the same physical machine. Modern hypervisors use hardware-assisted virtualization, virtualization-specific hardware, primarily from the host CPUs.
• Process virtual machines are designed to execute computer programs in a platform-independent environment.

Some virtual machines, such as QEMU, are designed to also emulate different architectures and allow execution of software applications and operating systems written for another CPU or architecture. Operating-system-level virtualization allows the resources of a computer to be partitioned via the kernel. The terms are not universally interchangeable.

Definitions

A "virtual machine" was originally defined by Popek and Goldberg as "an efficient, isolated duplicate of a real computer machine." Current use includes virtual machines that have no direct correspondence to any real hardware. The physical, "real-world" hardware running the VM is generally referred to as the 'host', and the virtual machine emulated on that machine is generally referred to as the 'guest'. A host can emulate several guests, each of which can emulate different operating systems and hardware platforms.

System virtual machines

The desire to run multiple operating systems was the initial motive for virtual machines, so as to allow time-sharing among several single-tasking operating systems. In some respects, a system virtual machine can be considered a generalization of the concept of virtual memory that historically preceded it. IBM's CP/CMS, the first systems to allow full virtualization, implemented time sharing by providing each user with a single-user operating system, the Conversational Monitor System (CMS). Unlike virtual memory, a system virtual machine entitled the user to write privileged instructions in their code. This approach had certain advantages, such as adding input/output devices not allowed by the standard system.

As technology evolves virtual memory for purposes of virtualization, new systems of memory overcommitment may be applied to manage memory sharing among multiple virtual machines on one computer operating system. It may be possible to share memory pages that have identical contents among multiple virtual machines that run on the same physical machine, what may result in mapping them to the same physical page by a technique termed kernel same-page merging (KSM). This is especially useful for read-only pages, such as those holding code segments, which is the case for multiple virtual machines running the same or similar software, software libraries, web servers, middleware components, etc. The guest operating systems do not need to be compliant with the host hardware, thus making it possible to run different operating systems on the same computer (e.g., Windows, Linux, or prior versions of an operating system) to support future software.

The use of virtual machines to support separate guest operating systems is popular in regard to embedded systems. A typical use would be to run a real-time operating system simultaneously with a preferred complex operating system, such as Linux or Windows. Another use would be for novel and unproven software still in the developmental stage, so it runs inside a sandbox. Virtual machines have other advantages for operating system development and may include improved debugging access and faster reboots.

Multiple VMs running their own guest operating system are frequently engaged for server consolidation.

Process virtual machines

A process VM, sometimes called an application virtual machine, or Managed Runtime Environment (MRE), runs as a normal application inside a host OS and supports a single process. It is created when that process is started and destroyed when it exits. Its purpose is to provide a platform-independent programming environment that abstracts away details of the underlying hardware or operating system and allows a program to execute in the same way on any platform. 

A process VM provides a high-level abstraction – that of a high-level programming language (compared to the low-level ISA abstraction of the system VM). Process VMs are implemented using an interpreter; performance comparable to compiled programming languages can be achieved by the use of just-in-time compilation.

This type of VM has become popular with the Java programming language, which is implemented using the Java virtual machine. Other examples include the Parrot virtual machine and the .NET Framework, which runs on a VM called the Common Language Runtime. All of them can serve as an abstraction layer for any computer language.

A special case of process VMs are systems that abstract over the communication mechanisms of a (potentially heterogeneous) computer cluster. Such a VM does not consist of a single process, but one process per physical machine in the cluster. They are designed to ease the task of programming concurrent applications by letting the programmer focus on algorithms rather than the communication mechanisms provided by the interconnect and the OS. They do not hide the fact that communication takes place, and as such do not attempt to present the cluster as a single machine.

Unlike other process VMs, these systems do not provide a specific programming language, but are embedded in an existing language; typically such a system provides bindings for several languages (e.g., C and Fortran). Examples are Parallel Virtual Machine (PVM) and Message Passing Interface (MPI). They are not strictly virtual machines because the applications running on top still have access to all OS services and are therefore not confined to the system model.

History

Both system virtual machines and process virtual machines date to the 1960s and continue to be areas of active development. 

System virtual machines grew out of time-sharing, as notably implemented in the Compatible Time-Sharing System (CTSS). Time-sharing allowed multiple users to use a computer concurrently: each program appeared to have full access to the machine, but only one program was executed at the time, with the system switching between programs in time slices, saving and restoring state each time. This evolved into virtual machines, notably via IBM's research systems: the M44/44X, which used partial virtualization, and the CP-40 and SIMMON, which used full virtualization, and were early examples of hypervisors. The first widely available virtual machine architecture was the CP-67/CMS (see History of CP/CMS for details). An important distinction was between using multiple virtual machines on one host system for time-sharing, as in M44/44X and CP-40, and using one virtual machine on a host system for prototyping, as in SIMMON. Emulators, with hardware emulation of earlier systems for compatibility, date back to the IBM System/360 in 1963, while the software emulation (then-called "simulation") predates it. 

Process virtual machines arose originally as abstract platforms for an intermediate language used as the intermediate representation of a program by a compiler; early examples date to around 1966. An early 1966 example was the O-code machine, a virtual machine that executes O-code (object code) emitted by the front end of the BCPL compiler. This abstraction allowed the compiler to be easily ported to a new architecture by implementing a new back end that took the existing O-code and compiled it to machine code for the underlying physical machine. The Euler language used a similar design, with the intermediate language named P (portable).^[8] This was popularized around 1970 by Pascal, notably in the Pascal-P system (1973) and Pascal-S compiler (1975), in which it was termed p-code and the resulting machine as a p-code machine. This has been influential, and virtual machines in this sense have been often generally called p-code machines. In addition to being an intermediate language, Pascal p-code was also executed directly by an interpreter implementing the virtual machine, notably in UCSD Pascal (1978); this influenced later interpreters, notably the Java virtual machine (JVM). Another early example was SNOBOL4 (1967), which was written in the SNOBOL Implementation Language (SIL), an assembly language for a virtual machine, which was then targeted to physical machines by transpiling to their native assembler via a macro assembler. Macros have since fallen out of favor, however, so this approach has been less influential. Process virtual machines were a popular approach to implementing early microcomputer software, including Tiny BASIC and adventure games, from one-off implementations such as Pyramid 2000 to a general-purpose engine like Infocom's z-machine, which Graham Nelson argues is "possibly the most portable virtual machine ever created".

Significant advances occurred in the implementation of Smalltalk-80, particularly the Deutsch/Schiffmann implementation which pushed just-in-time (JIT) compilation forward as an implementation approach that uses process virtual machine. Later notable Smalltalk VMs were VisualWorks, the Squeak Virtual Machine, and Strongtalk. A related language that produced a lot of virtual machine innovation was the Self programming language, which pioneered adaptive optimization and generational garbage collection. These techniques proved commercially successful in 1999 in the HotSpot Java virtual machine. Other innovations include having a register-based virtual machine, to better match the underlying hardware, rather than a stack-based virtual machine, which is a closer match for the programming language; in 1995, this was pioneered by the Dis virtual machine for the Limbo language. OpenJ9 is an alternative for HotSpot JVM in OpenJDK and is an open source eclipse project claiming better startup and less resource consumption compared to HotSpot.

Full virtualization

Logical diagram of full virtualization

 

In full virtualization, the virtual machine simulates enough hardware to allow an unmodified "guest" OS (one designed for the same instruction set) to be run in isolation. This approach was pioneered in 1966 with the IBM CP-40 and CP-67, predecessors of the VM family.

Examples outside the mainframe field include Parallels Workstation, Parallels Desktop for Mac, VirtualBox, Virtual Iron, Oracle VM, Virtual PC, Virtual Server, Hyper-V, VMware Workstation, VMware Server (discontinued, formerly called GSX Server), VMware ESXi, QEMU, Adeos, Mac-on-Linux, Win4BSD, Win4Lin Pro, and Egenera vBlade technology.

Hardware-assisted virtualization

In hardware-assisted virtualization, the hardware provides architectural support that facilitates building a virtual machine monitor and allows guest OSes to be run in isolation. Hardware-assisted virtualization was first introduced on the IBM System/370 in 1972, for use with VM/370, the first virtual machine operating system offered by IBM as an official product.

In 2005 and 2006, Intel and AMD provided additional hardware to support virtualization. Sun Microsystems (now Oracle Corporation) added similar features in their UltraSPARC T-Series processors in 2005. Examples of virtualization platforms adapted to such hardware include KVM, VMware Workstation, VMware Fusion, Hyper-V, Windows Virtual PC, Xen, Parallels Desktop for Mac, Oracle VM Server for SPARC, VirtualBox and Parallels Workstation.

In 2006, first-generation 32- and 64-bit x86 hardware support was found to rarely offer performance advantages over software virtualization.

Operating-system-level virtualization

In operating-system-level virtualization, a physical server is virtualized at the operating system level, enabling multiple isolated and secure virtualized servers to run on a single physical server. The "guest" operating system environments share the same running instance of the operating system as the host system. Thus, the same operating system kernel is also used to implement the "guest" environments, and applications running in a given "guest" environment view it as a stand-alone system. The pioneer implementation was FreeBSD jails; other examples include Docker, Solaris Containers, OpenVZ, Linux-VServer, LXC, AIX Workload Partitions, Parallels Virtuozzo Containers, and iCore Virtual Accounts.

Machine code

From Wikipedia, the free encyclopedia

 

Machine language monitor in a W65C816S single-board computer, displaying code disassembly, as well as processor register and memory dumps.

 

Machine code is a computer program written in machine language instructions that can be executed directly by a computer's central processing unit (CPU). Each instruction causes the CPU to perform a very specific task, such as a load, a store, a jump, or an ALU operation on one or more units of data in CPU registers or memory. 

Machine code is a strictly numerical language which is intended to run as fast as possible, and may be regarded as the lowest-level representation of a compiled or assembled computer program or as a primitive and hardware-dependent programming language. While it is possible to write programs directly in machine code, it is tedious and error prone to manage individual bits and calculate numerical addresses and constants manually. For this reason, programs are very rarely written directly in machine code in modern contexts, but may be done for low level debugging, program patching (especially when assembler source is not available) and assembly language disassembly. 

The overwhelming majority of practical programs today are written in higher-level languages or assembly language. The source code is then translated to executable machine code by utilities such as compilers, assemblers, and linkers, with the important exception of interpreted programs, which are not translated into machine code. However, the interpreter itself, which may be seen as an executor or processor, performing the instructions of the source code, typically consists of directly executable machine code (generated from assembly or high-level language source code). 

Machine code is by definition the lowest level of programming detail visible to the programmer, but internally many processors use microcode or optimise and transform machine code instructions into sequences of micro-ops. This is not generally considered to be a machine code.

Instruction set

Every processor or processor family has its own instruction set. Instructions are patterns of bits that by physical design correspond to different commands to the machine. Thus, the instruction set is specific to a class of processors using (mostly) the same architecture. Successor or derivative processor designs often include all the instructions of a predecessor and may add additional instructions. Occasionally, a successor design will discontinue or alter the meaning of some instruction code (typically because it is needed for new purposes), affecting code compatibility to some extent; even nearly completely compatible processors may show slightly different behavior for some instructions, but this is rarely a problem. Systems may also differ in other details, such as memory arrangement, operating systems, or peripheral devices. Because a program normally relies on such factors, different systems will typically not run the same machine code, even when the same type of processor is used. 

A processor's instruction set may have all instructions of the same length, or it may have variable-length instructions. How the patterns are organized varies strongly with the particular architecture and often also with the type of instruction. Most instructions have one or more opcode fields which specifies the basic instruction type (such as arithmetic, logical, jump, etc.) and the actual operation (such as add or compare) and other fields that may give the type of the operand(s), the addressing mode(s), the addressing offset(s) or index, or the actual value itself (such constant operands contained in an instruction are called immediates).

Not all machines or individual instructions have explicit operands. An accumulator machine has a combined left operand and result in an implicit accumulator for most arithmetic instructions. Other architectures (such as 8086 and the x86-family) have accumulator versions of common instructions, with the accumulator regarded as one of the general registers by longer instructions. A stack machine has most or all of its operands on an implicit stack. Special purpose instructions also often lack explicit operands (CPUID in the x86 architecture writes values into four implicit destination registers, for instance). This distinction between explicit and implicit operands is important in code generators, especially in the register allocation and live range tracking parts. A good code optimizer can track implicit as well as explicit operands which may allow more frequent constant propagation, constant folding of registers (a register assigned the result of a constant expression freed up by replacing it by that constant) and other code enhancements.

Programs

A computer program is a list of instructions that can be executed by a central processing unit. A program's execution is done in order for the CPU that is executing it to solve a specific problem and thus accomplish a specific result. While simple processors are able to execute instructions one after another, superscalar processors are capable of executing a variety of different instructions at once.

Program flow may be influenced by special 'jump' instructions that transfer execution to an instruction other than the numerically following one. Conditional jumps are taken (execution continues at another address) or not (execution continues at the next instruction) depending on some condition.

Assembly languages

A much more readable rendition of machine language, called assembly language, uses mnemonic codes to refer to machine code instructions, rather than using the instructions' numeric values directly. For example, on the Zilog Z80 processor, the machine code 00000101, which causes the CPU to decrement the B processor register, would be represented in assembly language as DEC B.

Example

The MIPS architecture provides a specific example for a machine code whose instructions are always 32 bits long. The general type of instruction is given by the op (operation) field, the highest 6 bits. J-type (jump) and I-type (immediate) instructions are fully specified by op. R-type (register) instructions include an additional field funct to determine the exact operation. The fields used in these types are: 

   6      5     5     5     5      6 bits
[  op  |  rs |  rt |  rd |shamt| funct]  R-type
[  op  |  rs |  rt | address/immediate]  I-type
[  op  |        target address        ]  J-type

rs, rt, and rd indicate register operands; shamt gives a shift amount; and the address or immediate fields contain an operand directly.

For example, adding the registers 1 and 2 and placing the result in register 6 is encoded: 

[  op  |  rs |  rt |  rd |shamt| funct]
    0     1     2     6     0     32     decimal
 000000 00001 00010 00110 00000 100000   binary

Load a value into register 8, taken from the memory cell 68 cells after the location listed in register 3: 

[  op  |  rs |  rt | address/immediate]
   35     3     8           68           decimal
 100011 00011 01000 00000 00001 000100   binary

Jumping to the address 1024: 

[  op  |        target address        ]
    2                 1024               decimal
 000010 00000 00000 00000 10000 000000   binary

Relationship to microcode

In some computer architectures, the machine code is implemented by an even more fundamental underlying layer called microcode, providing a common machine language interface across a line or family of different models of computer with widely different underlying dataflows. This is done to facilitate porting of machine language programs between different models. An example of this use is the IBM System/360 family of computers and their successors. With dataflow path widths of 8 bits to 64 bits and beyond, they nevertheless present a common architecture at the machine language level across the entire line. 

Using microcode to implement an emulator enables the computer to present the architecture of an entirely different computer. The System/360 line used this to allow porting programs from earlier IBM machines to the new family of computers, e.g. an IBM 1401/1440/1460 emulator on the IBM S/360 model 40.

Relationship to bytecode

Machine code is generally different from bytecode (also known as p-code), which is either executed by an interpreter or itself compiled into machine code for faster (direct) execution. An exception is when a processor is designed to use a particular bytecode directly as its machine code, such as is the case with Java processors.

Machine code and assembly code are sometimes called native code when referring to platform-dependent parts of language features or libraries.

Storing in memory

The Harvard architecture is a computer architecture with physically separate storage and signal pathways for the code (instructions) and data. Today, most processors implement such separate signal pathways for performance reasons but implement a Modified Harvard architecture, so they can support tasks like loading an executable program from disk storage as data and then executing it. Harvard architecture is contrasted to the Von Neumann architecture, where data and code are stored in the same memory which is read by the processor allowing the computer to execute commands.

From the point of view of a process, the code space is the part of its address space where the code in execution is stored. In multitasking systems this comprises the program's code segment and usually shared libraries. In multi-threading environment, different threads of one process share code space along with data space, which reduces the overhead of context switching considerably as compared to process switching.

Readability by humans

Pamela Samuelson wrote that machine code is so unreadable that the United States Copyright Office cannot identify whether a particular encoded program is an original work of authorship; however, the US Copyright Office does allow for copyright registration of computer programs and a program's machine code can sometimes be decompiled in order to make its functioning more easily understandable to humans.

Cognitive science professor Douglas Hofstadter has compared machine code to genetic code, saying that "Looking at a program written in machine language is vaguely comparable to looking at a DNA molecule atom by atom."