ACL—see Access Control List
Access Control List (ACL)
A list used by administration software for authorizing and validating access to controlled resources.
agent
A software module that provides a programmatic interface to the processing of a controlling process. In a client/server application the agent is responsible for preparing and exchanging information on behalf of either a client application or server application.
API—see Application Programming Interface
Application Programming Interface (API)
A set of related programmatic function definitions, exposing calling convention and format, used to facilitate access to software not directly implemented in the program.
application protocol
A mutually understood, predefined, sequence of messages exchanged between two application entities.
application resource
Data accessed using a Resource Manager.
asynchronous communication
A method of communication used between application entities in which the sending entity is able to proceed with its own processing without waiting for a reply to its request message. Contrast with synchronous communications.
atomicity
The "all-or-nothing" property of a transaction that insures that all of the transaction aware resources participating in the transaction either commit their associated changes or roll them back to their state prior to beginning the transaction.
authentication
A security mechanism used by a system to ensure that an application component has a valid identity.
authorization
A security mechanism, used by a system, to permit or deny access to a resource controlled by the system. See Access Control List.
blocking condition
A condition encountered in an application resulting in that application halting execution until something happens, such as the completion of a send operation, a response received, a timer pop, and so on.
bracket
A programmatic demarcation used to indicate the beginning and end of a transaction.
channel
A path established between two communicating end points over which communications transmission can take place.
Client
A software module that gathers and presents data to an application; it requests services and receives replies. This term can also be used to indicate the requesting role that a server module assumes when it requests services of another server module (see server-to-server flows).
client/server
A request/response programming style characterized by the application being divided into two or more cooperating components. The cooperating components are known as clients and servers. Clients request work of servers. In some cases a server takes on the role of a client when it requests the services of another server module (see server-to-server flows).
commit
A software action associated with making changes to associated resources permanent. Contrast with rollback. See two-phase commit.
daemon
A process that is actively running and does its processing in the background.
data-dependent routing
A process of selecting a server based on the contents of data contained in the request message being processed.
data independence
A characteristic of software processing that allows data to be accessed and updated without knowing the underlying storage format or relative structure of that data.
data marshalling
The process of rendering arguments and data values into a byte-stream format and packaging that data into a buffer so that it is suitable for transmitting between heterogeneous nodes connected using a communications network.
data unmarshalling
The process of rendering data to a node's native data representation. Typically, the data would have been previously marshalled into a format suitable for transmitting over a network. (see data marshalling).
decryption
The process of rendering an encrypted message into an intelligible format. Contrast with encryption.
deferred communications
Software processing that two modules communicate such that both applications need not be active simultaneously. Typically, deferred communications are performed using message queues. Also see asynchronous processing.
Directory Service
Software processing that maps names to locations.
distributed application, distributed computing
Software processing characterized by an application being composed of two or more software components that are physically located on more than one computer. Also see client/server.
domain
A component part of an architectural hierarchy. Typically, the domain is associated with an administrative partition of a complex distributed system, name space, or application.
durability
The property of a transaction that guarantees the permanence of modified data upon the completion of commit processing.
dynamic configuration
The ability to change the definition of an application while that application is active.
encryption
The process of converting a message rendering into a form where its contents are unintelligible to unauthorized parties. Contrast with decryption.
event
The occurrence of a condition, state change, or activity, of interest to one or more software components.
event notification
The process of signaling one or more software components that an event has taken place.
fault tolerance
A characteristic such that an application continues to run despite failures in its software components or communication interfaces.
forwarding
A processing technique by which pipeline parallelism is implemented, where a message is sent on to another service that will complete the work and reply to the original requesting application.
function shipping
A processing technique whereby the processing of an application is segmented such that functions are located near the data they need to access. Requests to access data are sent to the associated function and only the results of the processing are sent back to the requestor.
gateway
A software module capable of exchanging information between two different application environments and or networks.
global transaction
A transaction, which spans more than one Resource Manager.
heterogeneous computer platforms
A collection of computers utilizing different data formats, different operating systems, different internal data representation (byte ordering), and so on.
implementation transparency
A method whereby a requestor of a service need not be aware of how the service processing actually occurs, letting re-implementation of the service without impacting the requestor.
IDL—see Interface Definition Language
Interface Definition Language (IDL)
A language used to describe data types and functions, typically used in an implementation of a Remote Procedure call (RPC).
IPC - see Inter-Process Communications
Inter-Process Communication (IPC)
Various system provided mechanisms by which modules in separate address spaces (processes) communicate within a single computer. Examples of IPCs include shared memory, semaphores, message queues, mapped memory files, mailslots, domain sockets, atoms, and so on.
isolation
The behavior of a transaction that insures that concurrent operations on the underlying data do not interfere with each other.
load balancing
A software process that assigns inbound service requests to instances of servers capable of processing the request service, in a manner to optimize throughput and response time.
local client
A client application that executes on the same physical computer as the server application that has been selected to process a given cooperative request.
local transaction
A transaction, which is only active and known to a single Resource Manager.
location transparency
A software characteristic that lets a resource, or service, be identified in such a way that its name implies no particular network address or physical location.
message
One or more packets of data communicated between one or more programs.
middleware
Connectivity software that lets components of an application inter-operate across a network, despite differences in underlying communications protocols, system architectures, operating systems, databases, and other application services. Portions of middleware address communications while other portions provide an environment in which applications execute. Middleware can take on the following different forms:
peer-to-peer
A cooperative processing environment where two communicating entities have similar processing capabilities.
phase one
The first phase of a two-phase commit. In this phase the resources modified during the transactions are prepared for phase two. Phase one culminates with the recording of the transactions decision to a log. This phase is under the control of the Transaction Manager. The Transaction Manager sends a prepare notice to the involved Resource Manager(s). The Resource Manager(s) write their updates to stable storage, and respond with an indication that it is prepared to commit or does not wish to commit.
phase two
The second phase of a two-phase commit which announces the transaction decision to the Resource Manager(s) involved in the transaction, triggering the Resource Manager(s) to release locks and making the results of the transaction visible to others desiring access to the managed resource(s).
post an event
A software operation that indicates that a processing condition of interest has occurred.
presentation service
A software layer that converts data being transmitted between heterogeneous computers. Presentation Service is usually associated with client applications allowing end user interaction with an associated software application or system.
proxy
A software component that exposes a simplified programmatic interface to a specific server application. A proxy presents a consistent, well-defined interface, used by other software modules to facilitate communications between the user of the proxy and its associated application. Using the interface exposed by a proxy helps to isolate and insulate its user from having to know explicit details of the protocol required to communicate with the associated application.
queue
A data structure that holds messages that will be processed by a software procedure.
recovery
Software processing that attempts to restore an application to a known state after a transaction, server, network, or computer has failed.
Remote Procedure Call (RPC)
A software programming technique that causes the execution of a non-local program to execute as if it were a call to a local program.
request buffer
A data buffer containing a request for a service originating from a client.
response buffer
A data buffer containing the results of a request for service.
request/response
A software processing style characterized by clients sending request buffers to obtain service from server modules. The requests result in the server modules replying in the form of a response buffer. Both the requesting clients and responding servers must comprehend their respective rolls to adhere to the application protocol.
Resource Manager (RM)
An administrative software component that maintains the state of application resources (the most common being databases). Most DBMS software assumes that the Resource Manager has responsibility for those databases it manages.
rollback
The activity that restores an application resource to its state prior to the initiation of a transaction involved in modifying the resource. The Resource Manager owning the modified application resource carries out-processing a rollback. In cases involving two-phase commit processing, the rollback of activities associated with a transaction are under the control of the Transaction Manager.
RPC—see Remote Procedure Call
self-describing buffer
A buffer that contains a description of its own format and contents; a message format in which the data type(s) and length(s) are known or can be determined without additional external information.
server
A software module, made up of a set of well-defined inputs, outputs and processing, capable of accepting requests from clients, proxies or other servers. In a request/response style of distributed processing, after a server completes its requested processing, the server returns a reply to the requesting client.
service
The generic name given to application processing available for request by a client, characterized as having well defined inputs, outputs, and processing. Servers provide one or more services to requesting clients.
stateless interaction
The interaction between a client and server where the service neither uses nor maintains any extra information about the client to process its request; all the data necessary to complete the work is in the client's request buffer. Stateless interactions are usually associated with the request/response style of distributed processing.
stateful interaction
The interaction between a client and server where the server keeps track of the processing progress for a given client interaction. Typically, the client would have more than one message exchange with the server to complete its required processing. Stateful interactions are usually associated with a conversation style distributed processing.
synchronous communication
A style of communication between application entities in which the sending entity waits for a reply to its message before resuming its own processing. Contrast with asynchronous processing.
time-independent communications
A style of communications whereby two software module communicate, however, both do not have to be active simultaneously. Typically deferred communications are performed using message queues. This description does not make any sense. Looks like you can have copied the description of "deferred communications."
thread
An execution unit that is independently dispatched or is the embodiment of a unique execution context. A thread is comprised of a uniquely identifiable execution sequence and the memory it manipulates. One or more threads run within the context of a process.
time-out
An event that occurs when processing takes longer than expected or configured. Time-out events are often used to satisfy blocked operations.
transaction
(1) a bracketed unit-of-work, comprised of one or more software requests, such that all operations must complete to consider the transaction complete.
(2) a discrete operation, dispatched on request from an initiating client.
transaction initiator
An application software component that begins a transaction.
transaction terminator
An application software component that ends a transaction.
Transaction Manager (TM)
An administrative software package that manages global transactions across multiple computers or Resource Managers.
Transaction Processing Monitor (TPM)
A service broker that uses a transaction request to dispatch associated services. The TPM interacts with the Transaction Manager and Resource Managers providing a framework of control and administration over the server-side application runtime environment.
Two-Phase Commit (2PC)
A software protocol used by a Transaction Manager to coordinate the commit processing of a Resource Manager.
unit-of-work
A collection of one or more transactions that are considered autonomous. The unit-of-work is considered to be complete only with the sum of its parts has been completed.
XA Interface
An X/Open standard defining the protocol for communicating between an external Transaction Manager and a Resource Manager.
2PC—see Two-Phase Commit
|
Copyright © 2014 CA.
All rights reserved.
|
|