FI106989B - Method for writing and reading data - Google Patents

Description

106989

The present invention relates to a method for locating and reading data from a database in a rigid system in which a particular data set 5 is read from the database in chunks, after which the data set is compiled from separately read chunks, and in which system database data can change The invention further relates to a system comprising memory means, writing means for writing data to memory means, reading means for reading piece by piece of data stored in memory means, and buffering means for assembling a data set from reading pieces of reading means.

The invention relates to a database system in which only a limited amount of data can be read in one read transaction, whereby reading larger sets of data from a database requires reading pieces of information in increments. One example of such a system is the DX-200 telephone exchange system (of Nokia Telecommunications OY), which utilizes a TDMS database system. There is a limitation in the TDMS database system, which results in only about one kilobyte of data being transmitted at the interface of a single read or write transaction. However, because larger data sets need to be read from the database, these data sets must be read from the database piece by piece. Unless such a database system prevents writing data to a database for the duration of reading a given set of data, there is a risk that between reading pieces of the set the database information may change so that the read set may not be correct because some of its pieces have changed during reading.

• ·

Denying all write operations for the entire reading process is often -: **. · Not possible or meaningful as it significantly reduces the performance of the database system.

A solution is known in the art for verifying a piece-by-piece data set by calculating a checksum over the entire data set. If. ···. reading all pieces gives the same checksum, assuming that the data was successfully read. However, a major disadvantage of this known solution is that calculating checksums requires a lot of computational power. In addition, a change in data may not always appear in the checksum as an example, since two change operations may override each other for the purpose of calculating the checksum.

• »« · 2 106989

A solution is known in the art where, in connection with a data set, a time stamp is stored in the database indicating when the data set was last modified. In this case, the right of the fragmented data set can be checked by using a time stamp, that is, if the time stamp has not changed between read operations, the number is successful. The disadvantage of this known solution is that its reliable application requires the use of a frequently updated clock, especially in a system where writing operations can be performed very quickly. However, maintaining such a frequently updated clock requires a significant amount of computing capacity 10, because if the clock is updated too slowly, there is a risk that the data written by different writing operations will receive the same time stamp.

Further, a solution is known which utilizes two database-specific counters, the first of which is changed at the beginning of each write operation and the second is changed at the end of each write operation. In this case, when reading the data, it is first checked whether the values of the first and the second counter are the same. If this is not the case, then the data is being written to the database and reading will be interrupted. The disadvantage of this known solution is that it locks the entire database during write operations, that is, the data cannot be read even in a completely different part of the database 20 (a different set of data) when the current placement operation is targeted.

The object of the present invention is to solve the problem described above and to provide a more reliable and resource efficient solution for reading and writing data to a database. These objects are achieved by a method according to the invention, characterized in that information is provided. • ♦ ♦ '· "· the value of the counter of a given data set when writing data that:' ** modifies the data of that data set, checks the value of the data set • ·«: the value of the counter at the first reading of the data set and 30 when reading the last piece, and it is concluded that the data set ·: ··: has failed to read if the counter value has changed between these check times.

The invention further relates to a system in which the method according to the invention can be utilized. The system according to the invention is characterized by '*' · * 35 characterized in that the system includes data set-specific counters, that ·: ·: the writing means are adapted to change the values of the counters of the data sets to which the writes are written. the data is effective, and that the means for reading are adapted to check the value of the data set counter at reading the first piece of that data set and at the last reading of the data set and to indicate reading failure if the data set counter value has changed between said checks.

The invention is based on the idea of maintaining a separate transaction counter for each data set read from a database, and updating the counters of 10 data sets each time writing operations are performed on a database that affect those data entities. at start to its value at the end of reading 15. If the value of the counter has changed during the reading of the chunks, it means that the read dataset is invalid. The major advantages of the solution according to the invention are thus that it does not lock the entire database while writing data, provides more reliable information as to whether the data set has changed during reading and that very little resources are needed to check the reading success. or calculating specific checksums.

In a preferred embodiment of the method of the invention, the database is provided with a database-specific counter whose ar - ':, power is updated each time a write operation is performed on the database, followed by' · '/ 25 counters of the data sets affected new counter value. In this filter mode, data set-specific counters may consist of memory locations where the total counter value is stored, since they do not require actual counter functions. Thus, the database-specific counter provides a kind of clock function which can later be used to find out in what order the different data sets in the database have been updated.

. ···. Preferred embodiments of the method and system of the invention are disclosed in the appended dependent claims 2 and 4 to 5.

The invention will now be described in more detail by way of example with reference to the accompanying drawings, in which: FIG. 1 is a flow chart of a first preferred embodiment of the method of the invention, FIG. 2 is a block diagram of the system of the invention. of the first preferred embodiment, Figure 3 shows a flow chart of a second preferred embodiment of the method of the invention, and Figure 4 shows a block diagram of a second preferred embodiment of the system of the invention.

Figure 1 shows a flow chart of a first preferred embodiment of the method of the invention. The flow chart of Figure 1 can be utilized to control the reading and writing of database data in a database from which data sets must be read in pieces.

In block A, data set-specific counters are arranged in the system. For example, a single data set counter may consist of a memory location 15 where the number value stored is incremented by the writing process for each write event that writes data affecting that data set.

Block B checks whether the data is currently unread. That is, if a database is completely unreadable from the database, or alternatively one or more chunks of a data set are unread, proceed to block C. If unread, proceed to block J.

In block C, it is checked whether it is a reading of the first piece of a new data set. If so, proceed to block D.

In block D, when reading the first piece of en-25 of a new readable set, the value of the ··· ·; ** · counter for that data set is checked. For example, this value is stored in a specific memory location • · · * · “· where it can be read later.

• ·: * · * In block E, the next data set in turn is read; piece. In this case, of course, as much as 30 large pieces of data are read at a time, that is, in the case of a small data set, it can be read at once: · · ·: If, on the other hand, it is a larger database. ***: an entity that cannot be read at once, it will be read as quickly as possible * •. ri part.

After reading, in block F, it is checked whether the read piece was last 35 minutes. If it is found that this was the last piece of the data set, it proceeds to block G, where the computation value of the read data set is checked. The value of that counter is then compared with the value of the same counter at block one (block D). If it is found that the counter value has changed during the reading of the pieces belonging to the data set, proceeding from block H to block I, indicating that the reading has failed. This can be done, for example, by reading the data set as unread by the reading means, and then re-reading its beginning in the following possible situation.

In block J, it is checked whether data has been written. If there is no data to be written go to block K where the unwritten data is written to the data base 10.

When the data is written, it proceeds to block L, where the values of the counters of all data sets affected by the written data are updated. The counters that should be updated when writing certain data can be resolved, for example, by maintaining an index of 15 data sets for the database to which all data sets are associated with the data to be written to the database.

Fig. 2 is a block diagram of a first preferred embodiment of a system according to the invention. In the case of Fig. 2, the controller 1 is connected to the other parts of the system by means of three communication links L1 to L3. Controller 1 thus receives and transmits information stored in memory M over these communication links.

The data to be written to the database is written to memory M via the writing unit 2. The writing unit initially writes the received data to a database and then checks which data entities were affected by the data.

• · ·

If, by way of example, it is assumed that the writing unit 2 writes to the data M * in memory DATA1, it can be seen from Figure 2 that this information affects the data entities: 1, 2, and n. • · · v: Counters C1, C2, and Cn of entities by increasing their value by one.

30 Similarly, assuming that writer 2 writes data to DATA3 memory ·: **: M updates it to counters C1 and Cn because data DATA3 affects data set · *** · 1 and data set n. Data DATA3 can be, for example, < t subscriber system subscriber data, which affects, for example, the data of the groups (data set 1 and data set n) to which the «« · ··· * 35 subscriber belongs. Counters C1 to Cn may consist of, for example, memory locations, whereupon the location unit 2, when updating them, first checks their stored value and then stores a new value.

When reading data, the reading unit 3 checks the first value of the counter, 5 e.g. C2, of that data set when reading the first piece. The reading unit 3 then reads the data set piece by piece (as large pieces as possible). The reading unit 3 for each piece read is stored in memory. When the last piece of the data set in question has been read, and thus the data set is completely in the memory of the reading unit 3, the reading unit checks the counter C2 ar-10 of the read-backed data set. If it then detects that the value of the counter has changed between reading the first and last chunks, this means that the data set read has failed, in which case it starts to read the data set again from the beginning. If, on the other hand, the value of counter C2 has not changed, the reading is successful, whereupon the reading unit forwards the read data set to the controller 1 for further transmission via a communication link L1 to L3.

Figure 3 shows a flow chart of another preferred embodiment of the method according to the invention. The flow chart of Figure 3 is very similar to the flow chart of Figure 1. However, these flowcharts differ from each other in that, in addition to the data set-specific counters 20, the database-specific counter is maintained in connection with the flow chart of Figure 3.

Thus, when data is written to the database in block K ', the database-specific counter is then updated in block L', for example by increasing its value by one. This counter is therefore updated each time data is written to the database.

25 In block M ', the values of the computation of all data sets •; *: c are updated. The update is performed by · · «such that the new value of the database-specific counter (obtained in block L ') is: * · * counted as the counter value for all data sets to be updated. Thus • »« v; thus, the values of the counters of the data sets can, if desired, be determined by the order in which the data sets are updated.

·: ··: Figure 4 is a block diagram showing the operation of the system according to the invention. from the preferred embodiment. The block diagram in Figure 4 corresponds very well to •. 2 is a block diagram, but is different in that in the case of Figure 4, in addition to the data set-specific counters C1-Cn, the database-specific counters C are also utilized.

♦ «« · <· * »• ♦ 106989 7

When writing data to memory M ', the writing unit 2' initially updates the database-specific counter C, for example, by increasing its value by one. Subsequently, the writing unit 2 'updates the values C1 - Cn of the counters of all data sets affected by the data to be written to correspond to the new value of the 5 counters C specific to the database.

It is to be understood that the foregoing description and the accompanying drawings are merely intended to illustrate the present invention. Various variations and modifications of the invention will be apparent to those skilled in the art without departing from the spirit and scope of the invention as set forth in the appended claims.

«« I

1 1 «« <«« «i« i «i« i «i« i «i« i «i« i «i« i «« i «i« «i« i «i« i ««. • • • • • • • • •

Claims (5)

A method for entering data into and reading data from a database in a system, where a particular information unit is read bit by bit from the database, after which the information unit is compiled by the separately read bits, and in which system the data in the database can be changed between the reading of separate bits of the information unit, characterized by its own counters are arranged for the information units in the database, the value of a particular information unit's counter changes, when the said data is entered in the database which changes the information unit's data unit, the value of the information unit's relation to the information unit's relation bit and in connection with the reading of the last bit, and it is concluded that the reading of the information unit has failed if the value of the calculator has changed during the time between said control occasions. Method according to claim 1, characterized in that a database-specific counter is arranged adjacent to the database, and the value of the database-specific counter changes when new data is entered in the database, and the values in all of the information units' counters are changed to correspond to the new database-specific counter value. , which information units are affected by data entered in the database. *; * t 3. System belonging to memory means (M, M '), write means (2, 2') for entering data into memory means (M, M '), reading means (3) for bit reading of an information unit stored in the memory means (M, M '), and buffering means (3) for compiling the information unit of the bits read by the reading means, characterized in that the system includes information unit-specific counters (C1 - Cn), and that the writing means (2 , 2 ') are arranged, in connection with each entry to be performed in the memory means (M, M'), to change the values of the information integers (C1 - Cn) which are affected by the data entered by the writing entity, and The reading means (3) are arranged to check the value (C1) of the information. the data unit counter in connection with the reading of the first bit of information and in connection with the reading of the last bit and to indicate that the reading failed, unless the value (C1) in the information unit's counter has changed during the time between said checks. The system according to claim 3, characterized in that the system further comprises an integer counter (C), wherein the write means (2 ') are arranged to change the value of the integer counter (C) in connection with each write-in operation to be performed in the memory means (M'). ) and to change the values (C1) in the information unit's counters affected by data entered by the writing means (2 ') into the memory means (M1) 10 to correspond to the new value of the total counter (C). System according to claim 3 or 4, characterized in that said system is part of a telephone exchange system. «« «• · • ·« <«4 ····« · • · · • · · • · • · «·« «e« • «· ♦ 1 c. • · · m • ·« · · • · · · · · · · · · · · · · · ·