Computing and Computers - Batch Processing - BBC2 - 1980

Cirie's computing and computers we've come to the Royal Bank of Scotland to see how one of the major clearing banks uses its computers now computers are important in banking for example on an ordinary cheque there are numbers at the bottom of the cheque which are in a special computer format and the bank uses these numbers in order to identify individual accounts we ask Campbell Doherty the manager here about the importance of computers and how they have affected the way he deals with bank accounts computers are very important to bank managers nowadays and this due principally to the increase in the number of customers and the transactions passing through their accounts when I joined the bank at first all of these transactions were recorded in a ledger very laborious ly by a clock where he sat and posted the various entries debits and credits bringing out the resulting balances now of course all of the information is contained within the computer and it produces on our behalf and for the benefit of the customer statements on the transactions that they have carried out over say a period of a month such a statement as this where we see the various say debits that have come through the account the various credits and the resultant balances all this information debits credits and so on has to be held by the computer and not just for one customer the Royal Bank has about 500 branches throughout the country and information about customer accounts at all these branches is now held centrally in Edinburgh by the Royal banks computer system here at the computer center information is processed to keep customers accounts up to date and what we're going to do in this program is show you how the computer does it we have to start by seeing we're in the computer the information is stored here's our diagram of a computer's structure the large amount of data has to be kept in the part of the computer called backing store it's held in backing store until it's needed for processing the data is organized in files a file is a collection of data needed for some specific purpose there's usually more than one file in backing store but we're going to concentrate on the single file which holds data about customer accounts we spoke to the controller of systems development Norman Irvin we use 22 magnetic tapes to store the information about our customers accounts in the Royal Bank we have one and a half million accounts so therefore each of those tapes will contain approximately 70,000 accounts this state for example contains all the accounts for our branches in and around the Endura area but just what information is stored on this tape its information which is needed to produce a bank statement each line of the statement represents a single transaction and the whole statement is just one record in the bank's master file there's one record for each account in the master file let's see what a typical record might contain the toast item of data is the number this number identifies the record and it's based on the customers account number next there's the customers name and address and then details of transactions since the customers last bank statement for example these two transactions were checks on the 6th of December there was a check for four hundred and fifty pounds the number underneath is the number of the check the following day there was another check this time for four hundred and twelve pounds so there are four items of information on this record the number which identifies the record and then the named address and the two check transactions we call each of these items are filled so this record has four fields other records will have different numbers of fields because there are different numbers of transactions on each account when we put all the one and a half million records together we get the Master File and to refer to any particular record we use the number in the first field this numbers unique to that record in fact the records in the file are stored in sequence given by these identifying numbers a file ordered in this way is called a sequential file and maintaining the records in this sequence is vital for the way the file is brought up to date the method we use to update our Master File is batch processing the new transaction information is collected into a batch during the day and is then used to update the master file that evening so we can therefore say that the updating process takes place on a regular scheduled basis once every night and when we update this master file we don't physically change the file each night we create a brand new master file by copying the previous day's information from the master file and in the process we include the new transactions in the new master file record some records are copied from one type to another and in the process the new transactions are added let's see what happens to a typical record during this process file processing uses the computers working store so a copy of the record is made inside the working storm now if there's data about a new transaction this is added to the copy of the record these copies from the working store are used to build up a new master file but how does information for updating the file get into the computer most of the information comes from the bank itself when customers transfer money in and out of their accounts for example cashing a check just like saying that barbecue we've chosen to watch a Czech being cashed because this production effects only one customer's account their 40 frames in fire that burns in single space this transaction is to withdraw 50 pounds so what the computer must do is debit the account by 50 pounds to do this the chick is taken from the cash desk into the terminal room these terminals are connected to the bank's central computer and this is where the information is typed in here's our check and there's the debit for 50 pounds our transaction and many more transactions input during the day are stored ready for the batch update process later that evening it's now time for the operators to run the program which updates the Master File the program's name is caustic cau stands for current account update and the program begins by asking the operators to mount two tapes the first tape contains the old master file in the second is a blank tape for the new master file the program continues by displaying information for the operator to check if the information is correct the operator instructs the computer to begin the updating process to see how the updating process works on the complete master file we'll look at a simplified version we'll represent it on the left of the screen and on the right is a transaction file that's a temporary file containing all the transactions entered during the day the processing takes place in the working store will represent working store across the center of the screen we've arranged for the records in both files to be stored in order the first item to be processed is the first transaction that has to be joined on to the corresponding master file record so a copy of the first transaction is brought into the working store next we want a record from the master file we start at the beginning of the file and copy the first record into the working store here the transaction in working store is for the same account as the Master File record the computer can test this by checking that the identifying numbers match so the new transaction details are added on to the record and now the updated record can be used to start a new version of the master file that's the story for the first transaction now the next transaction can be processed copy it into the working store and also copy the next master file record but this time there's a difference the transaction isn't for this account the identifying numbers don't match so this record isn't changed it's copied directly onto the new master file the same thing happens with the third master file record it goes into the working store it doesn't match now for the fourth record into the working store and this time it does match so the new transaction is added onto the record and the updated record is copied to the new file and so it continues until all the records have been processed making a completely new master file this method taking transaction records one at a time and working through the master file in sequence there's a standard technique for processing a sequential file at high speed on average the computer can process about 70 records every second to achieve this the transaction file must therefore be in the correct sequence but transactions arrive at the computer center from our branches in random order also we have a great many transactions arriving from other sources for example yesterday we had to process about 175,000 cheques which had been presented for payment at other banks so what we do is we store all these transactions on magnetic disks just like this and then at about 5:00 p.m. after the date the collection process has ended for the day but before we start the master file updating run we have to get all these transactions sorted into account number sequence so what we do is we take all the transactions from the various disks we sort them and then store them in a single disk bank and this is the disk we use for the updating run that evening feature of batch processing if the transaction records were in a random order like this then the computer would waste time going backwards and forwards to find records that matched so instead the transaction file is sorted into the correct sequence then it's easy to work through the two files record by record with this technique we just start at the beginning of each file and work our way through to the end reading each record only once that's the essence of batch processing we can summarize what's happened in a system diagram which shows that transactions are first collected on several disks at this stage the transaction data is in random order next it's sorted and combined on a single disk this disk is used to update the master file held on magnetic tapes the updating process creates a new version of the master file and it also produces a large amount of printed output much of this printout consists of statements for bank's customers blank statement forms are fed into a line printer and it prints information about each account on to the forms then the statements are sent off to the branches but getting them there can cause problems as Norman Irvin explains each night we have to send reports and statements to our branches to arrive the following morning four branches in the end burning area there's no problem at all however we do have a problem for example in the case of careful in the Orkneys where the iboot has to be dispatched by 8:30 p.m. in evening to arrive in careful the following morning but of course the updating process is seldom completed before 11:00 p.m. each night so we get around this problem by taking advantage of the fact that the master file is contained in 22 separate magnetic tapes and we simply arrange for those branches farthest away from the computer center to have that accounts and the first tapes to be processed in this way we can produce our output reports in 22 separate batches according to the geographical replication of the branches this batch processing operation happens every night 22 times once for each tape in the master file and each tape is updated by copying is all this copying will be necessary you may think it's a waste of time to copy the records on our master file every night but we think this is the most cost-effective way of doing things and remember also as a bank we really are obliged to balance our books every night down to the last penny and this can be quite easily done as we produce the new master file each night and when we complete the updating process we don't throw away the old master file for data security reasons we keep several generations of our Master File so that if today for example we have some damage to our Master File we simply reconstruct the information from yesterday's master file and that's a bonus from batch processing because every day a new master file is created the previous version of the file is simply stored away but next day the new master file must itself be updated this means it becomes the input to the updating process and the output is yet another new master file and so it goes on day after day each newly created master file is the file which has to be updated the following day you

Channel: The Centre for Computing History

Views: 23,450

Keywords:

Id: fAselhq0Q8g

Length: 18min 57sec (1137 seconds)

Published: Mon May 30 2016