Robert K. Abbott and Hector Garcia-Molina. Scheduling real-time transactions: A performance evaluation. In VLDB, pages 1--12, 1988.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....5. Greater security problems Overview of RTDB Systems For the past decade, there has been a growing interest in the area of Real Time Database Systems (RTDB) The importance of research in this area has been emphasized in many workshops and conferences at the international fora in recent times [Son88, Son90, AGM92, Ulu92, Ulu94b, Bes96, BH95, BH96]. An RTDB system is an amalgam of a conventional real time system and a database system. In a real time system, the tasks have associated timing constraints, usually in the form of completion deadlines, and are scheduled in a way that they can be completed before their deadlines expire. Preserving ....

Robert K. Abbott and Hector Garcia-Molina. Scheduling Real-Time Transactions: a Performance Evaluation. ACM Transactions on Database Systems, 17(3), 1992.

....device is processed in priority order. Even at the network, it is easy to ensure that messages of high priority transactions are transmitted before those of other lower priority requesters. With respect to data access also, prioritybased pre emptive concurrency control algorithms such as 2PL HP [AGM88] and OPTWAIT [HCL90] have been developed. Removing priority inversion in the commit protocol, however, is not fully feasible. This is because, once a cohort reaches the prepared state, it has to retain all its data locks until it receives the global decision from the master this retention is ....

....also occurs at the priority of the transaction. The transaction priority assignment policy used is the widely used Earliest Deadline [LL73] transactions with earlier deadlines have higher priority than transactions with later deadlines. 5. 4 Concurrency Control The 2PL High Priority scheme [AGM88] is used for concurrency control of conflicting transactions. A concurrency control request from a high priority cohort for a data item held by a low priority cohort is satisfied by aborting the low priority cohort. If a low priority cohort requests a data item held by a high priority cohort, it ....

Robert K. Abbott and Hector Garcia-Molina. Scheduling Real-Time Transactions: a Performance Evaluation. In Proceedings of the 14th International Conference on Very Large Data Bases, August 1988.

.... Systems 7 Overview of RTDB Systems For the past decade, there has been a growing interest in the area of Real Time Database Systems (RTDB) The importance of research in this area has been emphasized in many workshops and conferences at the international fora in recent times [Son88, Son90, AGM92, Ulu92, Ulu94b, Bes96, BH95, BH96] An RTDB system is an amalgam of a conventional real time system and a database system. In a real time system, the tasks have associated timing constraints, usually in the form of completion deadlines, and are scheduled in a way that they can be completed before ....

Robert K. Abbott and Hector Garcia-Molina. Scheduling Real-Time Transactions: a Performance Evaluation. ACM Transactions on Database Systems, 17(3), 1992.

....device is processed in priority order. Even at the network, it is easy to ensure that messages of high priority transactions are transmitted before those of other lower priority requesters. With respect to data access also, prioritybased pre emptive concurrency control algorithms such as 2PL HP [AGM88] and OPTWAIT [HCL90] have been developed. Removing priority inversion in the commit protocol, however, is not fully feasible. This is because, once a cohort reaches the prepared state, it has to retain all its data locks until it receives the global decision from the master this retention is ....

....on behalf of a transaction also occurs at the priority of the transaction. The transaction priority assignment policy used is the widely used Earliest Deadline [LL73] transactions with earlier deadlines have higher priority than transactions with later deadlines. The 2PL High Priority scheme [AGM88] is used for concurrency control of conflicting transactions. A concurrency control request from a high priority cohort for a data item held by a low priority cohort is satisfied by aborting the low priority cohort. If a low priority cohort requests a data item held by a high priority cohort, it ....

Robert K. Abbott and Hector Garcia-Molina. Scheduling Real-Time Transactions: a Performance Evaluation. In Proceedings of the 14th International Conference on Very Large Data Bases, August 1988.

....not the low level transaction. ffl cannot be granted greater priority of execution over a transaction at a lower access class. There have been a number of papers in the real time databases literature that have explored priority based scheduling approaches with respect to conventional databases [1], 12] 14] The problem arises when these approaches are applied to secure databases, because covert channels can be introduced by priority based scheduling. All existing real time systems schedule transactions based on some priority scheme. The priority usually reflects how close the transaction ....

....at lower access classes. For example 2, the sequence of operations that would now be performed are rl 1 [x] r 1 [x] vwl 2 [x] vw 2 [x] dvwl 2 [y] vw 2 [y] c 2 rl 1 [y] r 1 [y] c 1 ru 1 [x] ru 1 [y] wl 2 [x] w 2 [x] wu 2 [x] wl 2 [y] w 2 [y] wu 2 [y] 4. 2 2PL High Priority In 2PL HP [1], all data conflicts are resolved in favor of the transaction with higher priority. When a transaction requests a lock on an object held by other transactions in a conflicting mode, if the requester s priority is higher than that of all lock holders, the holders are restarted and the requester is ....

[Article contains additional citation context not shown here]

Robert K. Abbott and Hector Garcia-Molina. "Scheduling Real-Time Transactions: A Performance Evaluation," ACM Transactions on Database Systems, Vol. 17, No. 3, pp 513-560, September `92.

....conflicts materialize. For RTDBMS, maximum concurrency (or throughput) ceases to be an expressive measure of performance. Rather, the number of timely commited transactions becomes the decisive performance measure [9] Most real time concurrency control schemes considered in the literature [1, 3, 31, 15, 29, 27, 28] are based on Two Phase Locking (2PL) which is a PCC strategy. Despite its Speculative Concurrency Control 145 widespread use in commercial systems, 2PL s long and unpredictable blocking times damage its appeal for real time environments, where the primary performance criterion is meeting time ....

Robert Abbott and Hector Garcia-Molina. "Scheduling real-time transactions: A performance evaluation." In Prooceedings of the 14th International Conference on Very Large Data Bases, Los Angeles, Ca, 1988.

....database: the more logically consistent schedules that the concurrency control allows, the more flexibility the database manager s scheduler has in determining a 3 logically consistent schedule that meets timing constraints. There has been significant work on scheduling real time transactions [AGM88, HS91, HLC90]; however, no widely applicable optimal scheduling algorithm has been found that will generate a logically consistent schedule that meets timing constraints if such a schedule exists. Thus, a real time concurrency control technique should allow as many logically consistent schedules as possible to ....

Robert Abbott and Hector Garcia-Molina. Scheduling real-time transactions: A performance evaluation. In 14th VLDB Conference, August 1988.

....scheduling policy for soft real time systems is Earliest Deadline First (EDF) 43] A deadline is some absolute time in the future when a response should be finished. A number of authors have found EDF to be a good policy for real time database systems, especially at low to moderate loads (e.g. [5, 68, 32, 34]) 3 Using the terminology of Daigle et al. a job is split up into a sequence of tasks. CHAPTER 2. BACKGROUND 20 Abbot and Garcia Molina in [5] report for systems with exponentially distributed interarrival times that EDF scheduling worked best at lower loads, and Least Slack, statically ....

....finished. A number of authors have found EDF to be a good policy for real time database systems, especially at low to moderate loads (e.g. 5, 68, 32, 34] 3 Using the terminology of Daigle et al. a job is split up into a sequence of tasks. CHAPTER 2. BACKGROUND 20 Abbot and Garcia Molina in [5] report for systems with exponentially distributed interarrival times that EDF scheduling worked best at lower loads, and Least Slack, statically evaluated, performed better at higher loads. For statically evaluated least slack, slack is measured once when a transaction arrives and is defined by: ....

Robert K. Abbott and Hector Garcia-Molina. Scheduling real-time transactions: A performance evaluation. ACM Transactions on Database Systems, 17(3):513--560, September 1992.

....systems [11] e.g. integration of real time and fault tolerance requirements, are of a much broader scope than actually addressed by traditional real time techniques such as processor scheduling. In particular, the need to consider accesses to shared (data) resources has drawn special attention [1]. In the non real time world, transactions [4] have been established to ensure consistency of shared data in spite of concurrent accesses and faults. They provide a powerful fault handling abstraction the so called All orNothing property (or failure atomicity property) that ensures that ....

....transaction models that replace the isolation property by more flexible concurrency control criteria [3] in order to achieve more concurrency and thus a higher throughput. In the same way, the isolation property has been identified in the real time world as a major obstacle to timely execution [1], i.e. isolating executions may cause deadlines to be missed. A vast amount of research has been investigated to make the mechanisms that implement the isolation property (pessimistically or optimistically) timeaware in the sense that they favor more urgent transactions with closer deadlines to ....

R. Abbot and H. Garcia-Molina, "Scheduling real-time transactions: A performance evaluation.," presented at 14th Conference on Very Large Data Bases,, Los Angeles, 1988.

....throughput of update transactions by 2 10 times[10] Xmas also supports group commit for the environment without a battery backup memory device. Two phase locking(2PL) protocol is used for the concurrency control of Xmas. The Priority inversion problem is resolved by priority inheritance protocol[1]. Xmas also limits the number of transactions executing concurrently to reduce the probability of aborting due to lock conflict. Xmas currently permits container level locking. To support finer granularity, a table in relational database can be horizontally partitioned into a number of ....

Robert Abbott and Hector Garcia-Molina. Scheduling Real-Time Transactions: A Performance Evaluation. In Proceedings of the 14th VLDB Conference, pages 1-- 12, 1988.

....the action. Using preemptive priority based thread scheduling[KSZ92] of Solaris 2.x, the current implementation platform of Xmas, higher priority transactions can be processed earlier than lower priority ones. The concurrency control protocol provided by Xmas is a form of two phase locking(2PL) AGM88] Priority inheritance and high priority protocols are used to resolve priority inversion problem. Xmas also limits the number of transactions executing concurrently to reduce the probability of aborting due to lock conflict. It has been suggested that very large lock granule is acceptable for ....

Robert Abbott and Hector Garcia-Molina. Scheduling Real-Time Transactions: A Performance Evaluation. In Proceedings of the 14th VLDB Conference, pages 1--12, Los Angeles, California, 1988.

....requirements. Thus, it should be a good scheduler for systems with soft deadlines and random or statistically determined arrivals and execution times. These properties are similar to those of real time database systems where Earliest Deadline First (EDF) scheduling has been investigated (e.g. [9, 10, 11, 12]) All the studies show that EDF scheduling works well at lower loads, allowing most of the responses to complete within their deadlines. We are most interested in these cases. During periods of overload, EDF suffers from the domino effect [13] whereby a response which is already late is given ....

Robert K. Abbott and Hector Garcia-Molina. Scheduling real-time transactions: A performance evaluation. ACM Transactions on Database Systems, 17(3):513--560, September 1992.

....can be run, with different time bounds, guaranteeing an upper bound on the time to produce some result. Mena [Mena82] has classified concurrency control algorithms into two classes: optimistic algorithms (e.g. Hari90b, Hari90a, Huan90, Kim91, Lin90, Son92] and pessimistic algorithms (e.g. [Abbo88, Stan88, Huan90, Sha91]) Bestavros has proposed (in [Best92] a new approach, Speculative Concurrency Control (SCC) which incorporates redundant computation into concurrency control algorithms. Speculative concurrency control redundantly executes the same transaction, under different ordering assumptions. The ....

Robert Abbott and Hector Garcia-Molina. "Scheduling real-time transactions: A performance evaluation." In Prooceedings of the 14th International Conference on Very Large Data Bases, Los Angeles, Ca, 1988.

....by removing data contention constraints so there is more flexibility in determining schedules that meet transaction timing constraints. In recent years, there has been a significant amount of research towards developing real time database models [18, 16] real time transaction scheduling [1], and real time concurrency control [21] However, there has been little work done on structuring and specification of real time transactions to better support data availability and timing constraints. Conventional transaction structures and specification techniques are not adequate for real time ....

....constraints. Other transactions are soft transactions which have some diminished value after a missed deadline. Transaction scheduling in a database system implies both transaction selection and concurrency control, which is a more complicated than typical task scheduling in an operating system [1, 13]. Concurrency control is one of the most widely studied areas in real time databases because many researchers have recognized the inherent conflict between meeting timing constraints and maintaining database consistency (for a survey of real time concurrency control and scheduling techniques see ....

Robert Abbott and Hector Garcia-Molina. Scheduling real-time transactions: A performance evaluation. In 14th VLDB Conference, August 1988.

....traditionally used 2PL HP for wide ranges of the system load for firm deadline transactions. 1 Introduction Database applications that require time constrained (real time) response to transactions are becoming quite common. Network Services Databases [SBSR93, Red86] and Stock market applications [AGM88, Voe87] are typical applications of real time database systems (RTDBS) Realtime transactions are the basic unit of work for RTDBS and transactions with deadlines have been categorized into hard, soft, and firm real time transactions. Transactions with hard deadlines have to meet their deadlines; ....

....transactions. In contrast to the above two, firm transactions are those which should not be considered once their deadlines are not met, as there is no value to completing the transactions after their deadlines. Typically, applications that have a definite window (e.g. stock market applications [Voe87, AGM88]) within which transactions need to be executed come under this category. There are several research issues [Gra92] on RTDBS of which the most active is developing transaction scheduling algorithms for soft and firm real time transactions. The scheduler of a RTDBS is responsible for assigning ....

[Article contains additional citation context not shown here]

Robert Abbott and Hector Garcia-Molina. Scheduling real-time transactions: a performance evaluation. In Proceedings of the 14th VLDB, pages 1--12. ACM, 1988.

....Real time concurrency control schemes considered in the literature could be viewed as extensions of either PCC based or OCC based protocols, whereby transactions are assigned priorities that reflect the urgency of their timing constraints. These priorities are used with PCC based techniques [AGM88, ACL87, SZ88, HSTR89, Sin88, SRL88, SRSC91] to make it possible for urgent transactions to abort conflicting, less urgent ones (thus avoiding the hazards of blockages) and are used with OCC based techniques [Kor90, HCL90b, HCL90a, HSRT91, KS91, LS90, SPL92] to favor urgent transactions when conflicting, less urgent ones attempt to ....

....T o u . CI u (t) X T i 2T u w i (t) Theta cv i u 3. If CI u (t) ff, then delay T o u s commitment, otherwise invoke the Commit Rule on T o u . 4 Performance Evaluation In this section, we present a comparative evaluation of the following protocols: 2PL with Priority Abort (2PLPA) [AGM88] as a representative of PCC based protocols, OCC BC [HCL90b] and WAIT 50 [HCL90a] as representatives of OCC based protocols, and SCC 2S and SCC VW as representatives of SCC based protocols. The RTDBS model that we used in our experiments consists of a multiprocessor DBMS operating on disk resident ....

Robert Abbott and Hector Garcia-Molina. Scheduling real-time transactions: A performance evaluation. In Prooceedings of the 14th International Conference on Very Large Data Bases, Los Angeles, Ca, 1988.

....scheduling [AGM89, KS91] In this paper, we focus on priority assignment of concurrent transactions. Concurrency control based real time database (time critical database) scheduling algorithms combine various properties of time critical schedulers with properties of concurrency control algorithms [AGM92, BMH89, CBB 89, HLC91, Sha88, SZ88, HSRT91] Priority scheduling without knowing the data access pattern is presented as a representative of algorithms with incomplete knowledge of resource requirements. The scheduling policies presented in [AGM92, HSRT91, HLC91, SZ88] combine either 2 phase ....

....properties of concurrency control algorithms [AGM92, BMH89, CBB 89, HLC91, Sha88, SZ88, HSRT91] Priority scheduling without knowing the data access pattern is presented as a representative of algorithms with incomplete knowledge of resource requirements. The scheduling policies presented in [AGM92, HSRT91, HLC91, SZ88] combine either 2 phase locking or optimistic concurrency control (OCC) with time critical schedulers. EDF HP (Earliest Deadline First with High Priority) LSF HP (Least Slack First with HP) EDF WP (EDF with Wait Promote) Virtual Clock, Pairwise Value Function [SZ88] are ....

[Article contains additional citation context not shown here]

Robert Abbott and Hector Garcia-Molina. Scheduling real-time transaction: Performance evaluation. ACM Transactions on Database Systems, 17(3):513--560, 1992.

....obtained from sensors, become inaccurate with the passage of time. Examples of such applications include autopilot systems, robot navigation, and program stock trading [13] While considerable work has been done on real time databases, most of it only assumes that transactions have deadlines [1, 7, 8, 10]. Database systems in which time validity intervals are associated with the data were discussed in [4, 16, 17, 9] Such systems introduce the need to maintain data temporal consistency in addition to logical consistency. The performance of several concurrency control algorithms for maintaining ....

Robert Abbort and Hector Garcia-Molina, "Scheduling Real-Time Transactions: A Performance Evaluation," ACM Transactions on Database Systems, Vol. 17, No. 3, pp. 513-560, September 1992.

....however, locking algorithms outperform OCC for non real time transac tions [AD85, ACL87, CS84] For soft deadline transactions, most of the research work has focused on priority assignment policy using locking based concurrency control. Several priority assignment policies have been proposed [AGM88, HJC93] However, extending 2PL with priority introduced priority inversion which is undesirable. This led to several priority based conflict resolution policies, such as Wait, Wait Promote (WP) and High Priority (HP) AGM88] schemes. In terms of performance, 2PL HP showed better performance ....

....control. Several priority assignment policies have been proposed [AGM88, HJC93] However, extending 2PL with priority introduced priority inversion which is undesirable. This led to several priority based conflict resolution policies, such as Wait, Wait Promote (WP) and High Priority (HP) AGM88] schemes. In terms of performance, 2PL HP showed better performance than OCC for soft RTDBS with finite resources [HCL90a] By definition, firm real time transactions are discarded when they miss their deadlines, as there is no value to completing them after they miss their deadlines. Unlike soft ....

Robert Abbott and Hector Garcia-Molina. Scheduling real-time transactions: a performance evaluation. In Proceedings of the 14th VLDB, pages 1--12. ACM, 1988.

....agreements (Noonan, 1990) and describes a requirement for explicit quality of service guarantees. This type of performance goals is statistical in nature, as they do not require that specific deadlines are met by individual units of work, which is the case with real time systems (Abbot and Garcia Molina, 1988). The transaction processing system is assumed to incorporate N nodes, S 1 ; SN . There are frontend nodes where user requests for the execution of multi transaction units of work arrive, coupled in a Shared Nothing architecture. Each transaction in the sequence of steps that makes up a ....

....the goal G p for the class. This policy requires no knowledge of the resource requirements of steps, taking into account only the goal for the workflow. This static assignement favors units of work with short response time goals, similarly to the earliest deadline first policy for realtime tasks (Abbot and Garcia Molina, 1988). PI : The priority of a transactional step of class WC p is set to 1 PIp , where P I p is the current value of the performance index for the class. The performance index P I p of class WC p is defined as P I p = R p G p ; where R p is the current estimate of average response time for ....

Abbot, R., and H. Garcia-Molina. 1988. Scheduling Real-Time Transactions: A Performance Evaluation.

....extending our results to soft and firm deadline systems by allowing for a profit loss past a transaction s deadline. Concurrency Admission Control 5 Control with forward validation (such as OCC BC [25] or SCC nS [5] or Pessimistic Concurrency Control (PCC) with Priority Abort (such as 2PL PA [1]) 2.1 Workload Admission Control Manager The source contains a set of transactions which are generated off line. Each enters the system at a random time and is first processed by the ACM. The decision of whether to admit or reject a transaction submitted for execution is based upon a feedback ....

....is the only conflict that T j has) 2PL has been criticized as being too pessimistic since it blocks transactions often unnecessarily and for potentially unbounded time. Real time Concurrency Admission Control 9 variants of 2PL have been suggested. One such variant is 2PL High Priority (2PL HP) [1] which augments 2PL with a priority based conflict resolution mechanism. A higher priority, lock requesting transaction aborts and restarts all lower priority, lock holding transactions which have a lock on the desired object in a conflicting lock mode. Moreover, 2PL HP prevents deadlocks due to ....

Robert Abbott and Hector Garcia-Molina. Scheduling real-time transactions: A performance evaluation. In Proceedings of the 14th International Conference on Very Large Data Bases, pages 1--12, Los Angeles, Ca, 1988.

.... In this paper we study our admission control mechanism in conjunction with two types of concurrency control protocols, namely Optimistic Concurrency Control with forward validation (such as OCC BC [29] or SCC nS [7] or Pessimistic Concurrency Control (PCC) with Priority Abort (such as 2PL PA [3]) 2.1 Workload Admission Control Manager The source contains a set of transactions which are generated off line. Each enters the system at a random time and is first processed by the ACM. The decision of whether to admit or reject a transaction submitted for execution is based upon a feedback ....

....commits or aborts (assuming this is the only conflict that T j has) 2PL has been criticized as being too pessimistic since it blocks transactions often unnecessarily and for potentially unbounded time. Real time variants of 2PL have been suggested. One such variant is 2PL High Priority (2PL HP) [3] which augments 2PL with a priority based conflict resolution mechanism. A higher priority, lock requesting transaction aborts and restarts all lower priority, lock holding transactions which have a lock on the desired object in a conflicting lock mode. Moreover, 2PL HP prevents deadlocks due to ....

Robert Abbott and Hector Garcia-Molina. Scheduling real-time transactions: A performance evaluation. In Proceedings of the 14th International Conference on Very Large Data Bases, pages 1--12, Los Angeles, Ca, 1988.

....materialize. Most real time concurrency control schemes considered in the literature and used in commercial systems combine Two Phase Locking (2PL) which is a PCC strategy, with a priority scheme to guarantee that the more urgent transactions are not blocked out waiting for less urgent ones [1, 26, 15, 24, 22, 23]. Despite its widespread use, 2PL has some properties such as the possibility of deadlocks and long and unpredictable blocking times that damage its appeal for real time environments. This led to a large body of research on alternatives to 2PL for RTDBS [16] This work has been partially ....

Robert Abbott and Hector Garcia-Molina. Scheduling real-time transactions: A performance evaluation. In Prooceedings of the 14th International Conference on Very Large Data Bases, Los Angeles, Ca, 1988.

....ordering also exhibit this problem [AJK 97] There has been very little work on addressing this important problem in multilevel secure databases. A number of papers in the real time databases literature have explored priority based scheduling approaches with respect to conventional databases [AGM92, SLL92] The problem arises when these approaches are applied to secure databases, To close all covert channels, a very high performance penalty must be paid. One suggested approach to improving the real time performance is to trade off mutual information transfer allowed by a covert channel with ....

Robert K. Abbott and Hector Garcia-Molina. Scheduling real-time transactions: A performance evaluation. ACM Trans. on Database Systems, 17(3):513--560, September 1992.

....However, because our technique was designed for real time applications, it is more important to measure temporal consistency than it is to measure throughput. To measure temporal consistency of transactions we examined the percentage of transactions that miss their deadlines (deadline miss ratio) [HSTR89, AGM88]. To measure the temporal consistency of the data we calculated the percentage of method requests that returned temporally invalid data to its transaction (temporal inconsistency ratio) Son92] Deadline Miss Ratio Temporal Inconsistency Ratio DL1: Vary Method Invocations TI1: Vary Method ....

Robert Abbott and Hector Garcia-Molina. Scheduling real-time transactions: A performance evaluation. In 14th VLDB Conference, August 1988.

....constraints, maximum concurrency (or throughput) ceases to be an expressive measure of performance. Rather, the number of transactions completed before their set deadlines becomes the decisive performance measure [Buch89] Most real time concurrency control schemes considered in the literature [Abbo88, Agra87, Stan88, Huan89, Sing88, Sha88, Sha91] are based on Two Phase Locking (2PL) which is a PCC strategy. Despite its widespread use in commercial systems, 2PL has some properties such as the possibility of deadlocks and long and unpredictable blocking times that damage its appeal for real time environments, where the primary performance ....

Robert Abbott and Hector Garcia-Molina. "Scheduling real-time transactions: A performance evaluation." In Prooceedings of the 14th International Conference on Very Large Data Bases, Los Angeles, Ca, 1988.

....In the RTSORAC model and prototype, we concentrate on soft real time database management, where providing predictability is desirable, but not necessary. The addition of timing constraints on transactions and data requires advanced transaction scheduling and concurrency control techniques [AGM88, BMHU89, YWLS94] Transaction schedules should meet timing constraints and they should maintain the logical consistency of the data in the database. Most conventional database system concurrency control techniques seek to maintain logical consistency of data while not supporting temporal ....

....constraints expressed in RTSORAC transactions into real time POSIX priorities for transaction processes. This mapping is designed so that the transaction process priorities realize Earliest Deadline First (EDF) scheduling. EDF scheduling has been shown to be effective in real time databases [AGM88] but implementing EDF scheduling using the capabilities specified by the POSIX interface is non trivial. The problem is that optimal EDF scheduling requires infinite priorities (one for each possible deadline) while POSIX mandates a minimum of only 32 priorities 1 . Furthermore, POSIX ....

Robert Abbott and Hector Garcia-Molina. Scheduling real-time transactions: A performance evaluation. In 14th VLDB Conference, August 1988.

....constraints, maximum concurrency (or throughput) ceases to be an expressive measure of performance. Rather, the number of transactions completed before their set deadlines becomes the decisive performance measure [Buch89] Most real time concurrency control schemes considered in the literature [Abbo88, Agra87, Stan88, Huan89, Sing88, Sha88, Sha91] are based on Two Phase Locking (2PL) which is a PCC strategy. Despite its widespread use in commercial systems, 2PL has some properties such as the possibility of deadlocks and long and unpredictable blocking times that damage its appeal for real time environments, where the primary performance ....

Robert Abbott and Hector Garcia-Molina. "Scheduling real-time transactions: A performance evaluation." In Prooceedings of the 14th International Conference on Very Large Data Bases, Los Angeles, Ca, 1988.

.... These attempts have been successful in the sense that they improved the performance of the basic PCC and OCC algorithms in the context of real time database management systems (RTDBMS) Most real time concurrency control schemes considered in the literature are based on TwoPhase Locking (2PL) [Abbo88, Stan88, Huan90, Sha91] a PCC algorithm that has been well studied in traditional database management systems (DBMS) Despite its widespread use, 2PL has some properties (such as the possibility of deadlocks and or long, unpredictable blocking times) which damage its appeal for RTDBMS, where in addition to ....

Robert Abbott and Hector Garcia-Molina. "Scheduling real-time transactions: A performance evaluation." In Prooceedings of the 14th International Conference on Very Large Data Bases, Los Angeles, Ca, 1988.

.... These attempts have been successful in the sense that they improved the performance of the basic PCC and OCC algorithms in the context of real time database management systems (RTDBMS) Most real time concurrency control schemes considered in the literature are based on TwoPhase Locking (2PL) [Abbo88, Stan88, Huan90, Sha91] a PCC algorithm that has been well studied in traditional database management systems (DBMS) Despite its widespread use, 2PL has some properties (such as the possibility of deadlocks and or long, unpredictable blocking times) which damage its appeal for RTDBMS, where in addition to ....

Robert Abbott and Hector Garcia-Molina. "Scheduling real-time transactions: A performance evaluation." In Prooceedings of the 14th International Conference on Very Large Data Bases, Los Angeles, Ca, 1988.

....materialize. Most real time concurrency control schemes considered in the literature and used in commercial systems combine Two Phase Locking (2PL) which is a PCC strategy, with a priority scheme to guarantee that the more urgent transactions are not blocked out waiting for less urgent ones [Abbo88, Stan88, Huan92, Sing88, Sha88, Sha91]. Despite its widespread use, 2PL has some properties such as the possibility of deadlocks and long and unpredictable blocking times that damage its appeal for real time environments. This led to a large body of research on alternatives to 2PL for RTDBS [Huan91] This work has been partially ....

Robert Abbott and Hector Garcia-Molina. "Scheduling real-time transactions: A performance evaluation." In Prooceedings of the 14th International Conference on Very Large Data Bases, Los Angeles, Ca, 1988.

....reflect the urgency of their timing constraints. These priorities are used in conjunction with PCC based techniques to make it possible for more urgent transactions to abort conflicting, less urgent ones (thus avoiding the hazards of blockages) Examples include the Priority Abort (PA) technique [3], Priority Inheritance (PI) technique [56] and variations of these techniques [55, 32, 7, 58, 64] These priorities are also used in conjunction with OCC based techniques to favor more urgent transactions when conflicting, less urgent ones attempt to validate and commit (thus avoiding the hazards ....

.... when conflicting, less urgent ones attempt to validate and commit (thus avoiding the hazards of restarts) Examples include the Broadcast Commit (BC) technique [40, 27] and the Wait 50 technique [26] Performance evaluation studies of these concurrency control techniques can be found in [3, 29, 57, 31, 1, 28, 30]. Other priority driven real time concurrency control protocols, which are not direct extensions of PCC or OCC, were also suggested in the literature. In [35] Kim and Srivastava studied and evaluated the potential performance improvement of using several protocols based on multipleversion ....

Robert Abbott and Hector Garcia-Molina. Scheduling real-time transactions: A performance evaluation. In Proceedings of the 14th International Conference on Very Large Data Bases, pages 1--12, Los Angeles, Ca, 1988.

....ceiling protocol, we reduce worst case blocking and context switches and improve system schedulability. There have been many previous attempts to adapt concurrency control techniques to soft or firm real time systems (RTS) scheduling, in which guarantees of schedulability are not required 2 [AGM88, AGM89, CJL89, HCL90a, HCL90b, 1 The term semantic is used in this sense in the database concurrency control literature. Syntactic concurrency control techniques are those that make use of no information beyond the sequence of read and write operations on database elements. 2 Terminology ....

Robert Abbott and Hector Garcia-Molina. Scheduling real-time transactions: a performance evaluation. In Proc. 14th Intl Conf on Very Large Data Bases, pages 1--12, August 1988.

....transactions that miss their deadlines and mean lateness. 2 Previous work There are several classes of real time database (time critical database) scheduling algorithms in which various properties of time critical schedulers are combined with properties of concurrency control algorithms [AGM88a, AGM88b, AGM89, BMH89, CBB 89, Har91, Sha88, SRSC91, SZ88, HSRT91] Priority scheduling without knowing the data access pattern is presented as a representative of algorithms with incomplete knowledge of resource requirements. The works in [AGM88a, AGM88b, AGM89, HSRT91, Har91, SZ88] fall into this ....

....of concurrency control algorithms [AGM88a, AGM88b, AGM89, BMH89, CBB 89, Har91, Sha88, SRSC91, SZ88, HSRT91] Priority scheduling without knowing the data access pattern is presented as a representative of algorithms with incomplete knowledge of resource requirements. The works in [AGM88a, AGM88b, AGM89, HSRT91, Har91, SZ88] fall into this category. Earliest Deadline First (EDF) Least Slack First (LSF) priority assignment policy with High Priority (HP) concurrency control method and Conditional Restart with estimated execution time on the main memory resident databases were proposed in ....

[Article contains additional citation context not shown here]

Robert Abbot and Hector Garcia-Molina. Scheduling real-time transactions: a performance evaluation. In Proceedings of the 14th Very Large DataBase Conference, pages 1--12. ACM, 1988.

No context found.

Robert K. Abbott and Hector Garcia-Molina. Scheduling real-time transactions: A performance evaluation. In VLDB, pages 1--12, 1988.

No context found.

Robert Abbott and Hector Garcia-Molina. Scheduling real-time transactions: A performance evaluation. In Proceedings of the 14th International Conference on Very Large Data Bases, pages 1--12, Los Angeles, Ca, 1988.

No context found.

Robert K. Abbott and Hector Garcia-Molina. "Scheduling Real-Time Transactions: A Performance Evaluation," ACM Transactions on Database Systems, Vol. 17, No. 3, pp 513-560, September `92.

No context found.

Robert Abbott and Hector Garcia-Molina. "Scheduling real-time transactions: A performance evaluation." In Prooceedings of the 14th International ConferenceonVery Large Data Bases, Los Angeles, Ca, 1988.

No context found.

Robert Abbott and Hector Garcia-Molina. "Scheduling real-time transactions: A performance evaluation." In Prooceedings of the 14th International ConferenceonVery Large Data Bases, Los Angeles, Ca, 1988.

No context found.

Robert Abbott and Hector Garcia-Molina. Scheduling real-time transactions: A performance evaluation. In Proceedings of the 14th International ConferenceonVery Large Data Bases, pages 1--12, Los Angeles, Ca, 1988.

No context found.

Robert Abbott and Hector Garcia-Molina. "Scheduling real-time transactions: A performance evaluation." In Prooceedings of the 14th International ConferenceonVery Large Data Bases, Los Angeles, Ca, 1988.

No context found.

Robert Abbott and Hector Garcia-Molina. "Scheduling real-time transactions: A performance evaluation." In Prooceedings of the 14th International ConferenceonVery Large Data Bases, Los Angeles, Ca, 1988.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC