WO2006050550A1

WO2006050550A1 - Method for exchanging data

Info

Publication number: WO2006050550A1
Application number: PCT/AT2005/000457
Authority: WO
Inventors: Raimund Kirner; Peter Puschner
Original assignee: Technische Universität Wien
Priority date: 2004-11-15
Filing date: 2005-11-15
Publication date: 2006-05-18
Also published as: AT501854B1; AT501854A1

Abstract

The invention relates to a method for exchanging data (DatAB) between a first computer system (CompA) and a second computer system (CompB), said data (DatAB) being represented on the first computer system (CompA) in a first platform-specific data representation and on the second computer system (CompB) in a second platform-specific data representation. The term 'platform-specific data representation' refers to the binary representation of numerical data which is conditional on the used platform (processor and compiler). The conversion of data (DatAB) between the first and second data representation is carried out in both directions on the first computer system (CompA), whereby the information (InfB) on the second computer system (CompB) required by the first computer system (CompA) is completely calculated on the second computer system (CompB), said information (InfB) being required for conversion of the data (DatAB) in both directions of data exchange.

Description

PROCEDURE FOR THE EXCHANGE OF DATA

The invention relates to a method for exchanging data between a first computer system and a second computer system, wherein the data is displayed on the first computer system in a first platform-specific data representation and on the second computer system in a second platform-specific data representation, and the conversion of the data between the first and second data representation is performed in both directions on the first computer system, wherein platform-specific data representation designating the binary representation of numerical data caused by the processor and compiler used. This concrete definition of platform-specific data representation is essential for the problem and the solution by the present invention. Hereinafter, heterogeneous computer systems refer to those computer systems which have a different platform-specific data representation.

The two computer systems - computer system being, for example, a processor on which data is processed - can consist of completely different hardware and software structures. In particular, as explained below, if the second computer system, which is also referred to below as "target computer system" or "target", the present invention, the resource consumption, which for the conversion of the data into another Platformspezifische data representation is necessary is minimized. The first computer system will hereinafter also be referred to as "host computer system" or "host".

The term "platform-specific data representation" introduced above is intended to be explained in more detail below: In this document, this term is understood to mean the binary representation of numerical data which is caused by the computer platform used (processor and / or compiler) Parameters for this binary representation of numeric data include:

Total size of data field for storing data value (e.g., number of bits)

• Address alignment of the data field (ie, of which size the specific address of the data field always has to be a multiple) • Binary coding of numbers (eg, one or two's complement for integer values, or IEEE format for floating point numbers, etc.)

Geometric arrangement of the individual data bits within the data field (e.g., endianity (litüe versus big endian), placement of elements of composite data within the data field, etc.)

The problem of converting data for exchanging information between heterogeneous computer systems has been known for a relatively long time, since the platform-specific data formats of different computer systems and the programs running thereon typically differ from one another in certain details. In principle, there are two ways to deal with the problem of different data formats:

1. Exchange of data using a uniform intermediate format: Each computer system must then convert the data between the internal platform-specific format and the uniform intermediate format. The advantage of this method is that computer systems can exchange data without having mutual knowledge of the platform-specific data formats used. The disadvantage of this method is its lack of efficiency, as certain aspects of data representation often need to be converted unnecessarily.

2. Exchange of data via a platform-specific data format: Data are only converted if it is necessary because of the platform-specific data formats of the computer systems involved. The advantage of this method is that the data conversion can be realized more efficiently than in the first method, since unnecessary conversions are avoided. Basically, two variants are conceivable:

a. Symmetric: Each system converts the data if necessary only when receiving (or only when sending). Thus, both computer systems take over part of the necessary data conversion. This method can be realized, for example, by adding a format description of the platform-specific data representation at the beginning of the transmitted data. A disadvantage of this variant is that both computer systems are busy with the data conversion.

b. Asymmetric: A computer system converts the data into the platform-specific format of the target system before it is sent or after it has been received in. its own platform-specific format. In this method, the second computer system is completely relieved of the expense of data conversion. However, this presupposes that knowledge about the internal platform-specific data format of the other computer system is known in the computer system performing the conversion.

A data exchange method according to method 2.b is particularly suitable for target computer systems, in which the resource consumption, which is necessary for the conversion of the data, is minimized. A disadvantage of the previously known methods according to Method 2.b, however, is that the support of the platform-specific data representation of a specific target platform a prioή must be built into the host, i. that the host must already have the necessary platform-specific information to exchange data with a specific target. The support of a new target therefore requires an explicit a priori adaptation of execution logic (e.g., software) on the host.

It is an object of the invention to improve a data exchange method according to method 2.b, so that even for new targets such an adaptation of the execution logic (for example software) on the host computer system is not necessary.

This object is achieved by a method mentioned in the introduction in that, according to the invention, the platform-specific information about the second computer system required on the first computer system, which information is necessary for the conversion of the data for both directions of the data exchange, completely on the second computer system is calculated.

The above object is further achieved with a corresponding non-contemplated method with a corresponding target computer system and with a test environment as set out in the claims.

The invention relates to a resource-saving method for exchanging data between a host computer equipped with a relatively large number of available resources and a relatively scarce resource-equipped embedded system (target). The embedded system uses little computing time and storage space for data transmission of the exchanged data. The conversion of the data is performed on the host computer, whereby the required platform-specific information about the target is automatically calculated on the target and transmitted to the host computer.

At the same time, the host requires a priori information about the platform-specific data format of the target, since the target itself calculates the required information and transmits it to the host.

The host thus takes care of the conversion of the data into the respective platform-specific data formats. The target only has to determine certain characteristics of its own platform once before the first exchange of user data and send it to the host. These characteristics include, for example, endianity (endian or big endian, i.e., the order of bytes within a data word). Knowledge of the endianity is required to interpret more information.

An important aspect of such a data exchange method is portability to new platforms both on the host side and on the target side. The method of this invention takes this into account by implementing the method for data conversion on host and target in platform-independent form. Neither the execution logic of the host nor the execution logic of the target need to be changed if the respective other system has been migrated to a new platform.

In order to use a data exchange method, it is also important that it is compatible with the development tools that are available for a specific computer platform. The method of this invention takes this into account by not requiring any special development tools for the realization. The method can be realized in a platform-independent form.

The process can be implemented with standard system development tools. Preferably, according to claim 2, the method for calculating the information on the second computer system is realized in a platform-independent form. The required execution logic on the side of the target can thus be realized independently of the platform, whereby the host computer can communicate with different targets without changing its execution logic. Thus, no specially adapted development tools are necessary to realize communication based on the method. According to claim 3 it is provided that the platform-specific information, which is calculated by the second Comptttersystem, an endian flag for describing the byte order, in which data are stored on the second computer system contains. This displays the storage strategies Little Endian and Big Endian. The value of this endian flag is the first data value within the information, since it determines how the further values of the information regarding byte order are to be interpreted. Furthermore, information about the length or value ranges of the basic data types may be contained by the second computer system.

In order to avoid the problem of different internal address alignment of complex data types on host and target, it is provided according to claim 4 that in the data the values of objects of complex data types, such as structures, in a flat form, i. as a sequence of elements of basic data types contained in the objects of complex data types.

According to claim 5 it is further provided that the first computer system and the second computer system are connected via an electronic data channel and automatically before the first transmission of the data, the information is transmitted from the second computer system to the first computer system.

With the features of claim 6, namely that the connection between the flat data structure of the data and the complex data structures is carried out by the second computer system via its own translation table, it is achieved that the size of the execution logic necessary for this connection is independent of type and number of data elements in the data.

Furthermore, it is envisaged that the objects of complex data types used for the connection between the transmitted objects of basic data types and those used on the target and also the transformation steps necessary for the calculation of the information are completely specified as a platform-independent method.

These features make it clear that the inventive method continues to function without modification should a particular target computer system be replaced by another target having different hardware and / or software structures. The method can therefore be easily ported to different platforms. All in all, it can be said that the process does not require any special support from the development tools, whereby the portability of the process is also achieved in this regard. The conversion of the data on the side of the host computer system can be realized with the aid of the platform-specific information in a platform-independent form and only has to support the conversion of all basic data types between the two platforms.

In addition to the method described above, a computer system according to claims 8-12, the object mentioned at the outset is also achieved with a test environment according to claims 13, 14.

In this test environment, the memory structure provides the required indirect data accesses if data types are not passed directly as value parameters to the subroutine to be tested, allowing the data exchange method of the test environment to be used without modification on different platforms.

In the following the invention is explained in more detail with reference to the drawing. In this shows

1 information flow between host and target,

FIG. 2 shows schematically the memory concept at the target for conversion between communicated data and internal data, and FIG

FIG. 3 schematically shows the extension of the storage concept at the target in order to be able to use this data exchange method for a decentralized test environment.

Fig. 1 illustrates the data flow and information flow between host (CompA) and target (CompB). It should be noted that before data (DatAB) can be exchanged between host (CompA) and target (CompB), the target (CompB) once transferred to the host (CompA) information (InfB) about platform properties of the target (CompB) ¬ must. These platform properties are automatically determined on the target (CompB) by means of short calculation steps. As soon as the information (InfB) has been received by the host (CompA), the host (CompA) has sufficient data available to know in which form the target (CompB) processes the data (DatAB). The data (DatAB) can be transmitted in both directions between host and target. The necessary conversion of the data format of the data (DatAB) is performed entirely on the host side (CompA). Claim 3 describes the information content of the information (InfB). 2 shows the memory concept on the target (CompB), by means of which objects of combined data types (DatB) are transmitted as a sequence of objects of basic data types (DatAB).

Basic data types are data types which consist of only one numerical value. Claim 4 describes this splitting of the objects of complex data types (DatB) into their elements of basic data types (DatAB). Depending on the software development environment used, different ranges of values as well as different number representations such as integer values or decimal values are available. Compound data types can consist of basic data types as well as of hierarchically composed data types. Typical forms of this composition, such as are supported in programming languages, are arrays (indexable sequence of similar data types), structures (association of any data types). The objects of basic data types (DatAB) are therefore called a flat data structure, while the composite data types of data types (DatB) are referred to as a complex data structure.

Since there are a great many different ways in which this composition of data types can be mapped concretely to memory cells, objects of combined data types are transmitted as a sequence of objects of the basic data types. The assignment of elements of the composite data types to the transmitted basic data types is possible via a table (TabB). This table (TabB) contains an entry for each object of a basic data type of (DatB), this entry containing a reference to the memory location where this object is expected to be executed in the target (CompB). Of course, this memory location can also be located within an object of a composite data type (DatB).

Reference is now made to Fig. 3, in which an extension of the memory interface of the data exchange method for realizing a decentralized test environment is described. The aim of this decentralized test environment is to run a (sub) program on the target computer system (CompB), the corresponding test data being provided by the host (Comp A) and transmitting the test result back to the host (Comp A) becomes. It may happen that the subroutine to be tested can not be called directly with the values of the test data, which is the case, for example, if the subroutine to be tested accesses the test data indirectly via other data structures In the case, these additional data structures must be set up and initialized before the test. This extension for the realization of a decentralized test environment is described in claims 13 and 14.

Exemplary realization of the data structures

In the following, it will be clarified by means of an example how the memory areas described for the data exchange method and a test environment based thereon can be realized and used. The data exchange method can be implemented both in hardware and in software. However, since programming languages allow execution logic to be very compact and in a generally known format, the programming language ANSI C has been chosen to describe certain mechanisms. The concepts of the test environment are also sketched for the ANSI C programming language.

Suppose it is the subroutine on the Target CompB

typedef struct {int a [3]; char b; } Sn_t;

void test (Sn_t * snl [20], Sn_t sn2);

to test. In order to demon- strate the flexible possibilities through the memory area (MGlue), it is assumed that the relevant test data consists of the two data sets svl and sv2 of the type Sn_t. It should also be noted that with respect to the input parameters of the subroutine to be tested, the first element of the array snl should point to the test value svl and the second element should be initialized as a zero pointer. The second input parameter is to be initialized directly with the test value sv2.

In order to be able to store the test data locally on the target (CompB), the following memory area is to be allocated:

Sn_t svl, sv2;

Since the procedure to be tested uses a more complex data structure in the first argument, additional memory area (MGlue) for allocating the complex data structure (DatB) must be allocated:

Sn t * av [20]; which is initialized with the following values:

av [0] = av [0] = &svl;

av [l] = (Sn_t *) 0;

The procedure to be tested could now be called as follows:

test (av, sv2);

To receive the test data coming from the host (Comp A), the following translation table (TabB) can now be used on the target (CompB):

typedef struct {void * ref; int size; } data_list_t;

data_list_t data [] = {\

{& (svl.a), sizeof (svl.a)}, \

{& (svl.b), sizeof (svl .b)}, \

{& (sv2.a), sizeof (sv2.a)}, \

{& (sv2.a), sizeof (sv2.b)}};

In the target (CompB), a receiving unit is implemented which interprets the objects of basic data types coming from the host (CompA) as a byte stream and distributes them to the corresponding memory cells which are specified in the table (TabB). The memory layout of objects of composite datatypes on Host (CompA) and Target CompB is structured differently, in a resource-saving way, by subtracting on the target (CompB) the data (DatB) with the translation table (TabB) on the objects of basic datatypes in ( DatAB).

This example demonstrates how to realize data communication from host (CompA) to target (CompB) for a remote test environment. The described data structures (table (TabB) and memory area (MGlue)) have been specified specifi cally for the (sub) program to be tested. The implementation of data structures for a data exchange from Target (CompB) to Host (Comp A) follows the same pattern.

Exemplary calculation of platform properties

In the following, the possibility of realizing the automatic calculation of the platform properties (information (InfB)) on the side of the target system (CompB) will be described. The examples used are merely intended to clarify the principle and do not represent a restriction on the scope of protection as defined in the claims.

There are several options for when the platform-specific information (InfB) is calculated. In order to save as many resources as possible on the target (CompB), it would be conceivable, for example, to run the calculation of the information (InfB) as an independent program on the target (CompB) and the platform-specific information (InfB) on the host system (CompA) for later communication with the target (CompB). An alternative possibility would be to include the calculation of the platform-specific information (InfB) directly in the application code of the target (CompB), which, however, occupies additional bytes of program code.

The platform properties described in the platform-specific information (InfB) contain in the first place the endianity (litfcte endian or big endian, i.e. the arrangement of the bytes within a data word). Knowledge of the endianity is required to interpret more information. Further additional information which may be contained in the information (InfB) are, for example, the value ranges of the basic data types.

The endianity could be calculated using the following platform-independent code in ANSI C:

int e = 1;

if (* ((char *) & e) == 1) endian = LITTLE; endian = BIG;

The value ranges of integral basic data types could be calculated with the following program code in ANSI C:

sizeof (char), sizeof (unsigned char), sizeof (short), sizeof (unsigned short),

sizeof (int), sizeof (unsigned int),

sizeof (long), sizeof (unsigned long)

For the basic data types for floating-point numbers, the calculation behaves similarly. The length of these basic data types can be calculated analogously:

sizeof (float), sizeof (double)

In addition, however, these FBsdata data types have internal partitioning into sign bit, exponent field, and mantissa field. A compiler for ANSI C BO / IEC 9899 would already float the data for this partitioning as constants in the file. have predefined h. In the case of older compilers, however, this partitioning data can also be calculated automatically.

Vienna, the

Claims

A method for exchanging data (DatAB) between a first computer system (CompA) and a second computer system (CompB), wherein the data (DatAB) on the first computer system (CompA) in a first platform-specific data representation and on the second computer system (CompB ) in a second platform-specific data representation, and the conversion of the data (DatAB) between the first and second platform-specific data representation in both directions on the first computer system. (CompA) is carried out, wherein platform-specific data representation denotes the binary representation of data which is caused by the used processor and / or compiler of the respective computer system (CompA, CompB)

characterized in that

the information (InfB) required on the first computer system (CompA) via the second computer system (CompB), which information (InfB) is required for the conversion of the data (DatAB) for both directions of the data exchange, is completely calculated on the second computer system (CompB) becomes.

2. The method according to claim 1, characterized in that the method for calculating the information (Inf B) on the second computer system (CompB) is realized in a platform-independent form.

3. The method according to claim 1 or 2, characterized in that the information (InfB), which is calculated by the second computer system (CompB), an endian flag for describing the byte order, in which data on the second computer system. (CompB) are stored contains.

4. The method according to any one of claims 1 to 3, characterized in that in the data (DatAB) the values of objects of complex data types (DatB) of the data on the target, such as structures, in a flat form, ie as a sequence in the complex Data types contained basic data types are stored.

5. The method according to any one of claims 1 to 4, characterized in that the first computer system (CompA) and the second computer system (CompB) are connected via an electronic data channel and automatically before the first transmission of data (DatAB) the information (InfB ) is transferred from the second computer system (CompB) to the first computer system (CompA).

6. The method according to claim 4 or 5, characterized in that of the second computer system (CompB), the connection between the transmitted data with flat data structure (DatAB) and the objects of complex data structures (DatB) via a translation table own (TabB) is performed.

7. The method according to any one of claims 1 to 6, characterized in that the necessary for the transformation between the transmitted data (DatAB) and the objects of complex data types (DatB) and the calculation of the information (InfB) onsschritte steps completely as a platform-specific method are specified.

8. Computer system (CompB) for a method according to one of claims 1 to 7, which is set up to receive the information (InfB) required for a first computer system (CompA) via the second computer system (CompB), which information (InfB) for Conversion of the data (DatAB) for both directions of the data exchange is necessary (InfB), to calculate completely.

9. Computer system (CompB) according to claim 8, characterized in that the method for computing the information (InfB) on the second computer system (CompB) is implemented in a platform-independent form.

10. Computer system (CompB) according to claim 8 or 9, characterized in that the information necessary for the calculation of the information (InfB) resources can be subsequently reused for other tasks after the calculation of the information (InfB).

11. Computer system according to one of claims 8 to 10, characterized in that the information (InfB), which is calculated by this computer system (CompB), an endian flag for describing the byte order, in which data on this computer system ( CompB) contains.

12. Computer system according to one of claims 8 to 11, characterized in that _/ in the data (DatAB) the values of objects of complex data types, such as structures, in a flat form, ie as a sequence of elements of complex data types contained elements of basic data types, get saved.

13. Test environment based on the method according to one of claims 1 to 7, characterized in that on a second computer system (CompB), the (Sub) program to be tested is executed, wherein text vectors are received from the first computer system (CompA) and Test results are transmitted back to the first computer system (CompA), and wherein a memory structure (MGlue) is used, which realizes the connection of the test data (DatB) to a program interface (DatBInt) of the (sub) program to be tested.

14. Test environment according to claim 13, characterized in that the description and initialization of the memory structure (MGlue) with the two interfaces (DatB) and (DatBInt) is implemented at the source code level as a platform-independent method.