SE517355C2 - Monitoring method and device - Google Patents

Abstract

The present invention relates to a supervision scheme and arrangement for removing obsolete records in a database of a mobile communications system, such as a DECT based PSTN/CTM system, particularly in a database handling subscribers. The invention comprises assigning an upper load limit (ULL) and a lower load limit (LLL) to the database (103, 105), assigning a minimum active period (MIP) to the records for at least which period they should not be removed (112, 113), assigning a duration of inactivity indicator (DII) to each record, repeatedly updating the duration of inactivity indicators (DII) for the records so as to indicate the time they have been inactive (114, 115, 201-225), and in dependence of reaching the upper load limit (ULL) removing inactive records in order so that records that have been inactive longest are removed first until reaching the lower load limit (LLL) or until only records which have been inactive for a shorter time than the minimum active period (MIP) remain (301-333).

Description

25 30 517 355 0 The entry for a subscriber who is in the database whose mobile device is switched on but stationary and who does not make or receive any calls will be inactive but not out of date.

The problem is finding a procedure to identify the inactive items that are most likely to be out of date so that they can be removed from the database.

One solution is to set a timer for each record and when the timer expires, the corresponding record is removed from the database.

The timer will be reset each time the item is accessed or updated. One problem with this solution is a high processor load due to the maintenance and monitoring of many timers. The database is scanned and records are removed even if the database has not reached a critical load level.

Other methods are based on the FIFO (First in First Out) order, where 'the oldest entry, or the. post soul has' been inactive for the longest period, _ is removed or exceeded when the database load becomes critical. This solution could lead to load problems - once the database load becomes critical, it will always be critical. Data must be requested and re-requested when an item is accessed or updated.

OBJECT OF THE INVENTION The object of the invention is to find a network node and a device for removing obsolete records in a database in a communication system, especially in a database which handles subscriber data, such as a VDB (Visitor Database) in a DECT network. . 10 15 20 25 30 a o o | en An additional goal is to avoid the problems described OVaIl.

This is achieved by a monitoring method and apparatus, which includes specifying an upper load limit (ULL) and a lower load limit (LLL) for the database, specifying a minimum active period (MIP) for the records where they cannot be removed at least during this period, ( DII) indicating an inactivity duration indicator for each record, repeatedly updating the inactivity duration indicators (DII) of the records to indicate the time they have been inactive, and depending on reaching the upper load limit (ULL) remove inactive records in turn , with items that have been inactive for the longest time first, until the lower load level (LLL) is reached or until only items that have been inactive for a shorter time than the minimum active period (MIP) remain.

Advantageously, the inactivity duration indicator of an item is initially, when it is created, set to zero and further always when it is accessed in a predetermined manner, for example by a subscriber. The inactivity duration indicator is further increased by one when a predetermined time has elapsed and if it has not reached s i t t maximum value.

The removal of items begins when the upper load limit is reached. A variable DIIn is set to the maximum value of the inactivity duration indicator. ' The database is scanned and all entries with an inactivity duration indicator equal to DIIn are removed, after which the variable DIIn is reduced by one and the database is scanned again, whereby all entries with an inactivity duration indicator equal to the new DIIn are removed. 10 15 20 25 o n | . Q »| . wo ø u u ø nu p ... 517 355 This step continues with successive decreases of DIIn until the lower load limit has been reached or when only items with an inactivity-duration indicator equal to or less than the minimum active period remain. The monitoring method and device in accordance with the invention effectively removes inactive items which are most likely to be reached at the top. obsolete when one should be found to be load limit The number of records removed depends on the lower load limit as well as a ndnn inactive period that an item should at least remain in the database.

An advantage of the invention is that it is dynamic and thus removes records when necessary.

One. A further advantage of the invention is that a certain amount of records are always available for use unless the database is poorly dimensioned or the upper load limit is set too close to the maximum capacity of the database.

BRIEF DESCRIPTION OF THE DRAWINGS The present invention will be better understood from the detailed description given below and the accompanying figures which are given by way of illustration only and are thus not limiting of the present invention.

Figure 1 shows database activity and status according to an embodiment of the present invention. lO l5 20 25 30 | n o o o n ø; u: u o u s .. n nu; Figure 2 schematically shows the cooperation between the monitoring device, the database, a CPU, I / O ports and other system blocks, respectively, according to the present invention.

Figures 3 and. 4 is a flow chart for initiating the monitoring method and apparatus of the present invention. for Figure 5 shows a flow chart for updating the DII records in the database.

Figures 6 show inactive items in accordance with the invention.

TERMS AND ABBREVIATIONS DII Inactivity duration indicator. a (0, l5), Advantageous number in the range where the value corresponds to a given time period that indicates how long an entry in the database has been inactive.

In-time Den. the minimum period during which an entry must be inactive before it can be removed from the database. This value is entered by the operator as a time period (YY mm dd / hh).

Level Monitoring level. Defines the length of time between updates in multiples of 3 and 24 hours. It is set per application and defined during start-up of the monitoring method. 10 15 20 25 30 LLL Lsup MIP Schedule 1 Schedule 2 T1 T2 Tsup 6 Lower load limit. The lower limit for acceptable memory load. When LLL is reached during database reset, the reset is stopped.

Load monitoring mechanism. This is used to notify when the critical memory load is reached, according to the ULL and LLL values specified by the operator. active DII value corresponding to (and Minimum period. calculated from) the I-time.

Update schedule indicating that the time between successive increases in DII will be three hours times the monitoring level when DII <8 and twenty-four hours times the monitoring level when DIII> 7.

Update schedule indicating that the time between successive increases in DII will be twenty-four hours times the monitoring level.

Period. T1 level * 3 hours level * 24 hours Time period. T2 Time monitoring mechanism. Tsup (T1) indicates that the time monitoring will initiate updating of DII at intervals of T1 and Tsup (T2) indicates that time monitoring will initiate updating of DII at intervals of T2 (advantageous at midnight). 10 15 20 un. «O. .nos-o ULL Upper load limit. The upper limit for acceptable memory load. When ULL is achieved, the recovery of the oldest data in the database is initiated.

DETAILED DESCRIPTION OF EMBODIMENTS The present invention proposes a monitoring method. and device for removing inactive records when the database reaches a critical defined upper memory load level as the load limit ULL.

The invention will make it possible to first remove items which have been inactive for the longest period of time. This will reduce the risk of removing items that are not yet obsolete.

The inventive database monitoring method and device will be controlled by an indicative time field, for example, on four bits, sonx will be added to each record in the database. This will be referred to as the inactivity duration indicator DII. DII will advantageously be numbered within the range (0, 15). A new record will always be assigned a DII = 0. The DII is updated repeatedly, ie. it is increased if it has not reached its maximum value, ie, 15. When a record is accessed, the DII is reset to zero. 10 15 517 355 Table 1. Ages on update schedule 1 and 2 records at monitoring level 1 as defined by DII.

Scheme 1 Scheme 2 DII Post Age, Post Age, Post Age, Post Age, min. max. my. max.

OO 3 hours O 1 day 1 3 hours 6 hours 1 day 2 days 2 6 hours 9 hours 2 days 3 days 3 9 hours 12 hours 3 days 4 days 4 12 hours 15 hours 4 days 5 days 5 15 hours 18 hours 5 days 6 days 6 18 hours 21 hours 6 days 7 days 7 21 hours 24 hours 7 days 8 days 8 1 day 2 days 8 days 9 days 9 2 days 3 days 9 days 10 days 10 3 days 4 days 10 days 11 days 11 4 days 5 days 11 days 12 days 12 5 days 6 days 12 days 13 days 13 6 days 7 days 13 days 14 days 14 7 days 8 days 14 days 15 days 15 8 days w 15 days w The invention can. adapted to be suitable for 'a large' variety of applications through the use of different At each level it is possible to use In table monitoring levels. two different update schemes; Scheme 1 and Scheme 2. 1 shows the different DII values and their corresponding record ages for the two schemes at monitoring level 1.

Using schedule 2, the DII is updated once a day, advantageously under low traffic loads, e.g. during the night.

Using Schedule 1, DII is updated every three hours until DII = 8 and after that every 24 hours. The maximum data storage periods are 15 days for schedule 2 and 8 days for schedule. 1 (poster soul have been inactive for longer than. These, periods will retain their maximum DII value and they will always be removed when the database when it critical memory load level). As the monitoring level increases, these update periods increase in multiples of these values, for example at level 3 the DIIs are updated every three days and every nine hours as shown in Table 2.

Table 2. Ages of update schedules 1 and 2 at monitoring level 3 as defined by DII.

Scheme 1 Scheme 2 DII Post Age, Post Age, Post Age, Post Age, min. max. my. max. 0 0 9 hours 0 3 day 1 9 hours 18 hours 3 days 6 days 2 18 hours 27 hours 6 days 9 days 3 27 hours 36 hours 9 days 12 days 4 36 hours 45 hours 12 days 15 days 5 45 hours 54 hours 15 days 18 days 6 54 hours 63 hours 18 days 21 days 7 63 hours 3 hours 21 days 24 days 8 3 days 6 days 24 days 27 days 9 6 days 9 days 27 days 30 days 10 9 days 12 days 30 days 33 days 11 12 days 15 days 33 days 36 days 12 15 days 18 days 36 days 39 days 13 18 days 21 days 39 days 42 days 14 21 days 24 days 42 days 45 days 15 24 days w 45 days w Genonl to increase DII more often at. a low value (<8) and. more rarely at higher values (> 7) as in the case of using Scheme 1 it is possible to set a very short MIP (minimum period during which an item must be inactive before it can and still get quite long maximums removed) data storage periods. 10 15 20 25 30 n o - ». I u u nu ~ n. n of n 517 355 lO Description. will in the following refer to the values applicable to level 1. The level of the monitoring scheme is as already defined defined for the application ...

The operator must enter values for the following parameters to activate the monitoring method and device according to the present invention. 0 The upper load limit ULL and the lower load limit LLL, respectively, which can be expressed as a percentage of the database capacity or as absolute numbers. 0 The MIP value, the DII value that corresponds to the minimum period during which an entry must be inactive before it can be removed from the database. Advantageously, the MIP value Lu is calculated for a specified I ~ time.

According to the present invention, records are removed from the database when ULL is obtained. The records are removed in turn with those with the highest DII values first. Removal continues until LLL is reached or until only records with DII S MIP remain.

An example of database activity and status according to the invention is presented in Figure 1. The upper half of the figure shows database activity and the lower half the status of the database.

Activity is indicated by solid arrows and status by dashed arrows. The DII values are updated according to schedule 1. For illustrative reasons, ULL is set to five entries and LLL is set to the I-time is selected to 15 hours, on three entries. which gives a MIP value of five. Initially, the database 11 contains four records POST 1-4, which have been inactive for different time periods, i.e., they have different DII values as indicated. A three-hour time period ends, which is detected by the time monitoring mechanism Tsup (3). The DII values are updated. After a while, a new record is added to the database with a DII value of zero (according to the inventive method), which is indicated in the resulting database 12. Another three-hour period ends. This is detected by Tsup (3) and the DII values are updated. The resulting database 13 with current DII values is displayed. It should be noted in particular that the DII values for POSTs 1 and 2 have not been increased yet because they exceed 7 and consequently, they are only increased when a twenty-four hour period ends. Another three-hour period ends, after which POST 3 is accessed in a way that resets the DII value to zero. Later, both a three-hour period and a twenty-four-hour period end at the same time, which is independently detected by the time monitoring mechanisms Tsup (3) and Tsup (24). Thus, the DII values of all records POST 1-5 are incremented by one as indicated in the resulting database 14. POST 4 is then activated and its DII value is reset. An additional three-hour DII update takes place, after which a database 15, as indicated, is obtained.

Another three-hour DII update occurs and an rqf entry is added.

At this point, the load monitoring mechanism Lsup detects that the current load, six items, exceeds the ULL value, five items, and a reset mechanism is activated.

First, all records with DII are removed by POST 2. 15. In our example, this corresponds to removal At this point, the load is compared with LLL and if the load exceeds LLL 10 15 20 25 30 »invention. ~. «. n 517 "355 12 the records with DII = 14 are removed as long as the current value of DII is higher than the specified MIP, which is the case in our example. However, there are no records with DII = 14, whereupon the value is further reduced At DII = 9, POST 1 is removed but still exceeds the LLL load, no more entries are to be removed until DII = 5 is reached.This is the same value as MIP and the reset mechanism is deactivated despite the fact that LLL is not reached.However, the database load is reduced Removed entries are displayed as crossed out in database 16.

Figure 2 schematically shows a complete system with interaction between the monitoring device. SA, the database. DB, a CPU, I / O ports and other system blocks, respectively, in accordance with the present invention. The main purpose of the monitoring device SA is as already described: 0 Time monitoring 0 Update of DII values (increase and reset) 0 Load monitoring 0 Removal of records The operator can change parameters, such as I-time, monitoring level, ULL and LLL, for monitoring schedule by I / The O-ports. The operator can be alerted.

Figures 3-7 show flow diagrams for implementing the monitoring method and the device according to the present Figures 3 and 4 show a flow diagram for initiating the monitoring method, Figure 5 shows a _flow diagram 10 15 20 25 30 u n u o c o a oo n o n n .o o u n n | c o .o A u 13 to update the DII values of the records in the database and Figures 6 and 7 'respectively show a flow chart to remove obsolete records.

In Figures 3-7, there are several boxes A, B, C, and D with the reference numerals 118, 304, 319 and 326, respectively, which indicate the flow of flow charts covering more than one figure and page, and consequently, they do not indicate any action. . Also note that for clarification reasons, a reference designation such as the direct preceding of a figure is written in italics in the text. The monitoring scheme SS is initiated 101. The administration Adm, ie. is requested I-time, level, ULL 103. the operator, with respect to values of the respective LLL, The operator specifies the values received by the inventive Monitoring Scheme 105. The schema then checks the format of the received values and that I-time> I o, level> o and ULL> LLL 1o7. if this is not the case, an error is reported to operator 108, who is also prompted to enter new parameter values.

T1 and T2 are calculated as follows 109: 0 T1 level * 3 0 T2 = level * 24 In addition, the following parameters are set 109: I Number_resets = 0 0 Status = PASSIVE 0 Update FALSE o Reset FALSE 10 15 20 25 30 ~ ø o Q noa Q o: o ø vu. u o .n p u n un | a u | un 517 355 14 Number_recovery indicates that the number of times the database has been restored by an inventive recovery macro 129. Status can be set to PASSIVE or ACTIVE. It is ACTIVE when, for example, the administration accesses the database and other procedures are thus prevented from being performed. The Refresh and Reset parameters are set to FALSE when the corresponding macros 121, 125, 201, 133, 301 are not running and are set to TRUE 'when run. lll will When the operator enters an I-time, which is less T2 update schedule 1 will be automatically selected. This is manifested through1 to set, a 'variable 2_hours on TRUE 113. Otherwise the variable 2_hours will be entered as FALSE 112 and schedule 2 will be selected.

The MIP value is calculated from the selected I-time and the update schedule used as follows. 0 MIP INTEGER (I-time / T1) If I-time 5 T2 113 0 MIP INTEGER (I-time / T2) if I-time> T2 112 MIP must always be less than 15. If the I-time specified by the operator results in a calculated MIP value greater than 14, the error is reported to the operator, not shown in the figures.

If a new update schedule is to be used, all records in the database must be scanned and their DII values changed according to the new schedule.

Time monitoring. Tsup, soul will initiate the increases of DII, started and the system Sys is informed. If scheme 1 has been chosen 10 15 20 25 30 u o | 1 and n n »4 o | .o n u 15 both Tsup (T1) and Tsup (T2) 115 are started and if schedule 2 has been selected, only Tsup (T2) 114 is started.

The load monitoring Lsup (ULL) is started 117. This continuously checks the database load and activates the database recovery macro 133 when ULL is reached. In addition, it handles the control of the load in the database against both ULL and LLL during the database recovery.

Onx there is no indication received. on Tsup (T1) 123, or Tsup (T2) 119, a macro update 121, 125, 201 is initiated. if Tsup (T2) is the one indicating.

The update macro is shown in Figure 5. The variable update is set to TRUE 203 and the status is checked 205. If status = ACTIVE, macro 207 waits until it receives status = PASSIVE. Then the first record is selected 210. It is checked whether the variable 2_timers is set to TRUE 211.

(X, Y) (0, l4) On: the answer is negative set the variables to 212. If the answer is positive, check if the variable U is set to zero 213. If the answer (X, Y) (X, Y) is positive is set to be (0.7) 215 and otherwise they are set to (8,14) 214. corresponds to the range of the DII values, which are to be increased. The DII of the first item is compared with the interval (X, Y) 217 and if it falls within the interval, the DII is increased by one.

Then the next item 222 is retrieved and its DII is incremented by one if it falls within the range. This continues until you get to the last entry 221. The Update variable is set to FALSE 223 and the macro ends 225. lO 15 20 25 30 ouuuuu un: I !! O U ut 10 ll 'Û. u uu u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u u. to be increased at three-hour intervals until entry. has been inactive for one day. Thereafter, it will be increased at twenty-four hour intervals, for example at 00:00, until DII reaches 15.

If the update schedule 22 is used, the DII for each inactive entry in the database will be increased at twenty-four hour intervals, for example at 00:00, for level 1. Once the DII reaches 15, it will not increase further.

When a new record is to be created or an old one is to be accessed or activated 127 a macro Inge Post is started. This macro, which is not shown in the figures, can be implemented in a known manner with an addition, namely that the DII value for added or accessed items is reset or set to zero.

If there is a request to store a new record in the database, it is full, SOm an alarm priority 3 will be triggered, indicating urgency or high priority.

The definition of an active record may differ between database applications. It does not have to be the case that every access to one. mail means that it is activated. For example, in a visitor database in a mobile telecommunication system, a subscriber record can be considered activated only when an action by the subscriber leads to the database being accessed. In this case, the subscriber record would be activated if it is accessed as a result of the subscriber making an outgoing call, but it would not be activated in this context if it is accessed as a 10 15 20 25 30 nu unc ø - nnn u. Nu Ü fi -Q \ J 04 OUT G fl 17 results of the operator providing the subscriber with an additional service.

When the load increases above the ULL, the Lsup (ULL) monitoring mechanism indicates that the load limit. is reached 131. This indicates a macro Reset 133, 301, which resets the database according to the flow chart in Figures 6 and 7. If status = ACTIVE 303, the macro 333 terminates. This occurs when the database is accessed by, for example, the administration. always such a variable Reset to TRUE 305.

The database load is checked after an access. If the status is PASSIVE, the Number_Resets is increased by one and one variable DIIn, which indicates that the current value of DII for which records to be removed is set to 15 307. The first record is retrieved 309 and it is removed 313 if its DII is the same as the current DIIn, ie. 311. The next record is retrieved 314 and removed 313 if its DII is equal to the current DIIn 311. This continues until the last record has 315. 317 and if the 333 is retrieved The database load is checked lower than LLL 321, the reset 331 is terminated, otherwise DIIn is reduced with a 323 and DDIn is compared with the MIP value 325. If they do not match, the records are retrieved again and all records with DII = DIIn, which is now 14, are removed 309-315. This procedure continues until the load is <LLL 321, ie. the reset is successful, or DIIn = MIP 325, ie. recovery failed - the load is not sufficiently reduced.

In the event that it fails, and if DIIn = MIP 325 and the load <: ULL 327, an alarm priority is issued 1. 329 indicating low priority. If DIIn = MIP 325 and the load is greater than or equal to ULL 327 then an alarm priority 2 328 is issued indicating average priority. In these cases, the operator 10 15 20 25 30 .S17 355 18 should take steps to adjust either the ULL and / or LLL values or the I-time. The Reset variable is then set to FALSE 331 and the macro ends 333.

The number of resets The number of resets that have taken place is saved and can be reported to the operator on request 135, together with a report on: the current ULL and I-LLL time values 139. The number of resets will then be reset 141.

If the operator sees that too many resets are taking place, it may be necessary to adjust the ULL, LLL or I-time values or to resize the database. This can be done through a macro Keys, which do not show the figures. The macro can allow the administration to change parameters such as level, I-time, ULL and LLL. Status is set to ACTIVE during this change, thus preventing other procedures from running. The monitoring schedule SS can receive a request from the administrator Adm to terminate the monitoring 155. Consequently, the load and time monitoring mechanisms 157, 159 are stopped.

The operator is notified that it has succeeded 161 and the inventive monitoring scheme is terminated 163.

The inventive monitoring scheme and the device can be used in a number of different areas such as, for example, databases that handle mobile subscribers, network topology and route selection databases.

The invention can in particular be implemented in a database such as an electronic mail box which handles electronic mail messages. The invention is intended to prevent the mailbox from becoming full. If e-mail messages are to be deleted, they should advantageously be stored in a storage medium, for example a hard disk.

Advantages of the present invention include a dynamic mechanism that only deletes records when necessary, i.e. when you reach an upper load level, and guarantees that you always find space for a new record provided that the database is properly dimensioned and that the upper load- too close to the database level is not set to be too high, ie. maximum capacity.

As the invention is thus described, it will be apparent that it may not be seen. All variations are to be varied in several ways. as exceeding the scope of the invention. such modifications 'soul would' be. obvious to a professional. is intended to be included within the scope of the following claims.

Claims (12)

10 15 20 25 30 517 355 = ~~ nu o: 20 PATENTKRAV A monitoring method for removing inactive items in a database in a communication system, in particular in a DECT (Digital Enhanced Cordless Telecommunications) based CTM (Cordless Terminal Mobility) wherein an inactivity (DII) system, for which (DII) duration indicator each item is specified, inactivity duration indicators are repeatedly updated for the items so that they indicate the time they have been inactive (114, 115, 201-225), characterized by the steps to: - specify an upper (ULL) and a lower 105), load limit load limit (LLL) for the database (103, - specify an active minimum period (MIP) for the items during at least which period they are not to be removed and (112, 113), - depending on the attainment of the upper load limit (ULL) remove inactive items in turn with records that have been inactive for the longest time only until the lower load limit (LLL) soul has been reached or until only records remain, inactive for a shorter time than the ndnst the active period (MIP) (301-333). A monitoring method according to claim 1, characterized in that it is implemented in a database for managing subscriber data. A monitoring method according to claim 1, characterized in that it is implemented in a database such as an electronic mail box_ with records consisting of electronic mail messages. A monitoring method according to claim 3, characterized by the storage of inactive electronic mail messages soul is removed from the electronic mail box in a storage medium. A monitoring method according to any one of claims 1-4, characterized by indicating the upper load limit (ULL) and the lower load limit (LLL) as a percentage of the maximum capacity of the database. A monitoring method according to any one of claims 1-5, characterized by setting the inactivity duration indicator (DII) initially to zero when the corresponding record is created and. also1 always when the entry is activated. A monitoring method according to claim 6, characterized by increasing the inactivity duration indicator (DII) when a predetermined time period (T1, T2) has elapsed and if it has not reached its maximum value (210-222). A monitoring method according to claim 6, characterized by increasing the inactivity duration indicator (DII) more often when it has a low value and more rarely when it has a higher value (210-222). 10 15 20 25 30 II I »av: IIo I I: I en: I I II II I I I co» I I I II II I c II I nu II I 1 II I n p I I n I I: I a .nu a I I I I on I; now co; u men I. I I o: I II I I I a a I I I I I I I - II of I. o; n vv a un. I 22 A monitoring method according to claim 6, characterized by increasing the inactivity duration indicator (DII) when three or twenty-four hours or multiples thereof have elapsed. A monitoring method according to any one of claims 6-9, characterized by not increasing the inactivity duration indicator (DII) if it has reached its maximum value. A monitoring method according to any one of claims 6-10, characterized by indicating the inactivity duration indicator (DII) as a four-bit field. A monitoring device for removing inactive items in a database in a mobile communication system, in particular in a DECT (Digital Enhanced Cordless Telecommunications) based CTM (Cordless Terminal Mobility) system, characterized by means for indicating an upper load limit (ULL ) and a lower 105), (MIP) load limit (LLL) for the database (103, - means to indicate a minimum active period for the records during which at least they are not to be removed (112, 113), - means to indicate an inactivity duration indicator (DII) for each record, - means for repeatedly updating the inactivity duration indicators (DII) of the records to indicate the time they have been inactive (114, 115, 201 ~ 225), and 23 depending on achievement of the upper load limit (ULL) means for removing inactive items in turn, where items that have been inactive for the longest period of time are removed only until the lower load limit (LLL) is reached or until only p cheeses remain that have been inactive for a shorter period than the minimum active period (MIP) (301-333). ono-u 1 - ~ vu-