AU2003253231B2 - A data processing system - Google Patents

Description

"A data processing system" INTRODUCTION 5 Field of the Invention The invention relates to data processing systems, and more particularly to initialisation and/or modification of a system to perform transactions. 10 Prior Art Discussion Many data processing systems which perform transactions require a large manpower input for initial programming and subsequent maintenance/upgrade. This is because of the highly complex nature of business requirements. For example, in the field of securities processing a system may handle many 15 different financial products, and for each product there are many functions to implement events (or "corporate actions") each comprising a number of transactions. As customer and financial needs and processes change and as regulations change, the corporate actions or events must be often modified and tested. Also, such complexity can lead to poor system performance for real time transaction processing because of the number of interactions between database and memory tables and retrieval of 20 multiple program functions. Watanabe, "Agent-oriented model for managing long-lived transaction, based on work-flow and task graph" (1999) describes a 3-layer transaction model and management method based on work-flow and task-graph. 25 As functions are typically programmed individually, albeit with some re-use of code, the initialisation and ongoing modification is often very time-consuming and expensive. The invention addresses these problems. 30 A reference herein to a patent document or other matter which is given as prior art is not to be taken as an admission that the document or matter was known or that the information it contains was part of the common general knowledge as at the priority date of any of the claims. 35 Where the terms "comprise", "comprises", "comprised" or "comprising" are used in this specification (including the claims) they are to be interpreted as specifying the presence of the stated features, W -aric'GARNODEL\737205 -plcd pas - 16 Ma 09 dc 2 integers, steps or components, but not precluding the presence of one or more other feature, integer, step, component or group thereof. SUMMARY OF THE INVENTION 5 According to the invention, there is provided a data processing system comprising a processor, a memory, an input interface, and an output interface, the processor being programmed to perform data processing operations to complete an event comprising a plurality of transactions each arising from a rule, wherein: 10 the system comprises a plurality of objects, each responsible for performing an event; each object comprises a container containing a series of at least one mask; 15 each mask comprises ON/OFF flags, each flag switching on or off an associated unit of executable code; said units of executable code: 20 being fixed, being stored separately from the object, being addressed by the object, and 25 implementing one of said transactions; the masks are stored in at least one word having a plurality of masks; 30 each object stores a numerical value representing the bits of each word, the word comprises binary bits and the numerical value is a decimal value; the masks are in a linear series flat logical structure, the processor being programmed to execute the executable code units of the masks in sequence; and 35 at least some of the transactions are asynchronous, without a return direction acknowledgement. WaricGABNODEL\737205 -repIacd page- 16 Mar 09 doc 3 In another embodiment, a flag is a binary bit. In a further embodiment, a word is a 64-bit long. 5 In one embodiment, the object contains an array comprising a plurality of indexes, each index addressing a word comprising a plurality of masks. In a further embodiment, the method comprises the steps of deciding in sequence if each of a limited 10 number of pre-stored units of executable code are required for an event, and setting a switch value in a mask for the event accordingly. In one embodiment, the switch values are set as attributes of an object representing the event. 15 In another embodiment, the flags are set with a fixed offset in a word for each unit decision. DETAILED DESCRIPTION OF THE INVENTION Brief Description of the Drawings 20 The invention will be more clearly understood from the following description of some embodiments thereof, given by way of example only with reference to the accompanying drawings in which: Fig. I is a diagram illustrating a processing structure for a data processing system of the 25 invention ; W:WarieGABNODEL\737205 - replad pge - 16 Mar 09 doc WO 2004/010290 PCT/IE2003/000105 -4 Fig. 2 is a more detailed diagram illustrating a rule array; and Fig. 3 is a flow diagram illustrating system initialisation. 5 Description of the Embodiments Referring to Fig. 1 a data processing system 1 is for executing operations for a financial institution. The business aspects of the operations may, however, be of any 10 type and the invention is concerned instead with the underlying technical features for system performance, ease of initialisation, and ease of maintenance/upgrade. The system 1 comprises a number of objects 2, only one of which is shown. Each object 2 is responsible for performing a business event comprising multiple 15 transactions. Each object 2 comprises: a container 3, a series of (in this case three) masks 4 contained within the container 3. Each mask 4 comprises four ON/OFF flags, each flag switching on or off an 20 associated unit of executable code 5. The units 5 of executable code are fixed and are separately stored, being addressed by the object 2. If an object 2 has only one mask 4, the complete event is implemented by executing some or all of the four units 5, each unit 5 implementing a transaction. The 25 maximum capacity of an object 2 is 16 masks, and so up to 64 transactions may be used to implement the event.

WO 2004/010290 PCT/IE2003/000105 -5 In this embodiment an event is called a "corporate action", being the term used for an event for a securities system. Such an event may, for example, by splitting of a shareholder's set of shares. 5 An important aspect is that there is only a limited number of executable units 6, in this embodiment four. These four units have code to implement the following transactions, expressed at the business level. (i) Keep Existing Securities Yes/No - does the customer keep their existing shares 10 as a result of this corporate action? If No then these securities are transferred out of the account. (ii) Receive New Securities Yes/No - does the customer receive new securities as a result of this corporate action? If Yes then the new securities are transferred into the account. 15 (iii) Receive Cash Yes/No - does the customer receive cash as a result of this corporate action? If Yes then a cash amount is transferred into the account. (iv) Pay Cash Yes/No - does the customer have to pay cash as a result of this corporate action? If Yes then a cash amount is transferred out of the account. 20 In this embodiment, each transaction is a single-direction asynchronous transfer, without an acknowledgement. It has been found that such a requirement for the transactions provides for efficient system operation and for effective breakdown of events. 25 Of course, the operations or transactions may be of any type other than the above, depending on the nature of the business. An advantageous aspect is that there is a limited and fixed number defined for the event.

WO 2004/010290 PCT/IE2003/000105 -6 In each mask 4 a "1" bit flag indicates that the associated unit 5 is to be used, and a bit "0" flag indicates that it is not to be used. Each executable unit 5 implements a transaction arising from a rule such as the rules (i) to (iv) set out above. There is therefore no need for inter-table fetches and so database access is kept to a minimum. 5 In practice, an object may have many masks 4. Thus, the masks 4 are stored in one or more longs, each of 64 bits, each long containing sixteen masks. Where there are more than sixteen masks there are two or more longs. 10 The structure is not logically hierarchical. The container 3 contains a number of masks, all masks being at the same level, in a "flat" logical architecture. The system processor executes the units 5 of the masks in succession, treating all of the masks and associated transactions as being at the same level. 15 The longs are stored in a "rulearray" 20 comprising, as shown in Fig. 2: an index 21 being the position of the long in the rulearray, a decimal value 22 being the value of the long at that position, and each long is represented at the binary level by 64 bits. 20 The binary representations of the decimal value is the series of bits making up the long. This manner of storing the masks is very efficient and simple - additional transactions being achieved by simply adding to the longs. Referring to Fig. 3, initialisation of an object for a corporate action is shown. In this 25 diagram the series of four decision steps on the right hand side sets the mask for each action, the commands: rule = rule + 1, 2, 4, 8 inserting the bit in the mask at the correct position. In this diagram the first position is the LSB, and so on.

WO 2004/010290 PCT/IE2003/000105 -7 When a mask is defined the long is updated to include the new mask. The "factor" sets the position in the long. For example, the third position is indicated by factor value 2. The command: 5 rule = rule + (factor * 10) in which (factor * 10), shifts the mask value to the correct position in the long. For example, for the third position in a long there is a factor value of 2 and (factor * 10) adds all preceding values to the mask value to inherently define the position. 10 In Fig. 3 the parameter "index" means the number of the long, the index being 0 for the first long. The rulearray is built up in iterations as shown in Fig. 3. Once set up it does not change until off-line modification is made by executing the method of Fig. 3. 15 This is a major advance for system initialisation and ongoing modification/upgrade, as it significantly minimises system architecture complexity and engineering manpower. All that is required for a new upgrade is to execute the process of Fig. 3 to create a new object. Modification of an existing object is achieved by simply 20 deleting the existing one and generating a fresh one using the method of Fig. 3 and generating a fresh one according to the process of Fig. 3. Also, system performance is excellent even for complex events because of the "flat" series mask sequence. This avoids iterative loops between different code hierarchical levels as has happened in the prior art. Another advantage arising from the structure of the objects 2 is that 25 overall system architecture is easy for systems analysts to visualise and understand. It will be appreciated that the invention allows a system to have a simple structure even if the operations to be performed are complex. Re-configuration is easily achieved by simply re-executing the setup process of Fig. 3. 30 WO 2004/010290 PCT/IE2003/000105 -8 The invention is not limited to the embodiments described but may be varied in construction and detail. For example, the events implemented by the objects may be other than those referred to as "corporate actions". Also, the masks may have flags other than binary bits, although binary representation is an efficient one.

Claims (9)

1. A data processing system comprising a processor, a memory, an input interface, and an output interface, the processor being programmed to perform data processing operations to complete 5 an event comprising a plurality of transactions each arising from a rule, wherein: the system comprises a plurality of objects, each responsible for performing an event; each object comprises a container containing a series of at least one mask; 10 each mask comprises ON/OFF flags, each flag switching on or off an associated unit of executable code; said units of executable code: 15 being fixed, being stored separately from the object, 20 being addressed by the object, and implementing one of said transactions; the masks are stored in at least one word having a plurality of masks; 25 each object stores a numerical value representing the bits of each word, the word comprises binary bits and the numerical value is a decimal value; the masks are in a linear series flat logical structure, the processor being programmed to 30 execute the executable code units of the masks in sequence; and at least some of the transactions are asynchronous, without a return direction acknowledgement. 35

2. A data processing system as claimed in claim 1, wherein a flag is a binary bit.

3. A data processing system as claimed in claims I or 2, wherein a word is a 64-bit long. W:narie'GABNODEL\737205 - replad pge- 16 Ma 09doc 10

4. A data processing system as claimed in any one of the preceding claims, wherein each object contains an array comprising a plurality of indexes, each index addressing a word comprising a plurality of masks. 5

5. A data processing system substantially as herein described with reference to the accompanying drawings.

6. An initialisation method for a system as claimed in any one of the preceding claims, the 10 method comprising the steps of deciding in sequence if each of a limited number of pre-stored units of executable code are required for an event, and setting a switch value in a mask for the event accordingly.

7. An initialisation method as claimed in claim 6, wherein the switch values are set as attributes 15 of an object representing the event.

8. An initialisation method as claimed in claim 7, wherein the flags are set with a fixed offset in a word for each unit decision. 20

9. A computer program product comprising software code for performing a method of any one of claims 6 to 8 when executing on a digital computer. W:.,,WcGABNODEL\737205 - replad pge. - 16 M, 09 dc