D.P. Anderson, \Meta-Scheduling For Distributed Continuous Media", Technical Report CSD-90-599, Computer Science Division, EECS Department, University of California at Berkeley, 1990.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....computes the resource allocation for the new application without touching admitted ones allocations. Further, Q RAM model does not consider precedence constraints among tasks of an application which is important in the context of real time systems. Continuous Media Resource (CM resource) model [2] is a theoretical framework that provides endto end performance guarantees to applications using continuous media (such as digital audio and video) Clients make resource reservation for worstcase workload. The meta scheduler coordinates with the CPU scheduler, network and le system and ....

D.P. Anderson, \Meta-Scheduling For Distributed Continuous Media", Technical Report CSD-90-599, Computer Science Division, EECS Department, University of California at Berkeley, 1990.

....CPU time can be reserved at the destination computer and the gateway between subnets 1 and 2 so that compression algorithms can be used to reduce the actual data rate going through subnet 2. 3. 5 Related work Our scheduler for a non reservable resource is different from Anderson s metascheduler [And93]. The metascheduler is used to coordinate multiple system components CPU, networks, file systems, 41 etc. and negotiate end to end guarantees on behalf of clients. It is resource reservation based. Our non reservable resource scheduler aims at the management of a single resource which is not ....

D. P. Anderson. Meta-scheduling for distributed continuous media. ACM Transactions on Computer Systems, 11(3), August 1993.

....time into the network. This time is compared with subsequent packets at all nodes, and the ff channel rate regulator maintains the original spacing between packets. A similar scheduling method provides utilization guarantees for a variety of resources including networks, CPUs, and file systems [3]. 97 The above systems are ideally designed for multimedia communication. Unfortunately these networks are not widely implemented. One of the strengths of the LAP scheduler is that it can operate on current existing networks that have no performance guarantees. If implemented on a real time ....

Anderson, D.P., "Meta-Scheduling for Distributed Continuous Media," UCB/CSD Tech. Rept. No. 90/599, October 1990.

....complexity simply the most relevant part of application development. A multimedia interface will have to take into account difficult problems such as paralellism, synchronization of data streams, real time constraints, asynchronous events, very large data objects, CPU demanding activities, etc. [5, 1]. And all those problems not only claim for better hardware and base software support, but for much more complex logic. The aim of this toolkit is to ease the process of multimedia interface composition in two ways: by isolating the programmer from low level details of multimedia interaction ....

....(such as image processing techniques) Jitter: When many graphical effects take place simultaneously, jitter begins to be noticeable. And it will become a real problem if software rendered video is used. A general mechanism of CPU reservation, overriding Unix s scheduling, will be necessary (see [1]) VCA and others: Graphic servers will benefit from the incorporation of image capture capabilities from cameras (and possibly scanners) new versions of ACPA drivers may incorporate MIDI; etc. Editors: A wealth of editors, particular of this toolkit, are urgently needed: depth, transparencies ....

David P. Anderson. Meta-scheduling for distributed continuous media. Technical report, Computer Science Division, EECS Department, University of California at Berkeley, Berkeley, CA 94720, October 1990.

....technique for dealing with the issue of real time data streaming is to reserve critical resources for a data stream. A case for resource reservation in distributed multimedia applications is made by Mosse (1995) who is also one of the few researchers to highlight the importance of fair billing. Anderson (1990) develops a formal framework for a reservation and scheduling mechanism for real time data traffic, which can be extended to networked environments end to end delivery also depends on the network and client (Anderson et al. 1992) A less formal approach to resource reservation and scheduling ....

Anderson, D. P, 1990. Meta-Scheduling For Distributed Continuous Media. Technical Report CSD-90-599, University of Berkeley Computer Systems Division.

....and display time. At the root node, setup time and display time are synonymous since the amount of time before the display of the entire presentation is exactly the setup time. Online monitoring of the presentation at runtime can be preformed by using techniques of meta scheduling as described in [And90] 14 To support the two message passing protocols, two methods are defined for each object class in the SRT. They are AcqTime and DisTime and can be invoked by sending acquiretime and distribute time messages respectively to the object involved. AcqTime method is invoked during the upward ....

D.P. Anderson. Meta scheduling for distributed continuous media. Technical Report 90-006, University of California, Berkeley, October 1990.

....is to define a burst parameter in addition to rate and frame size. The burst parameter specifies the maximum number of media elements that may arrive early that is, before they should arrive according to the regular arrival rate. The model of linear bounded arrival processes (LBAP) used in [Anderson 1993] defines the maximum number of messages in a stream during any time interval t as Rt B where R is the rate and B is the maximum size of burst. The advantage of using this model is that it nicely reflects the characteristics of multimedia sources: multimedia data read from disks is usually ....

Anderson, D.P. (1993), Meta-Scheduling for Distributed Continuous Media. ACM Transactions on Computer Systems, Vol. 11, No. 3.

....as the two alternative approaches and show why distributed multimedia database systems should be able to deal with both. The use of predictable data availability based on, e.g. operating system scheduling and bandwidth reservation techniques, respectively, has been studied previously, e.g. [4, 9, 25]. In particular, the use of dynamic quality of service control within the context of predictable data availability has been investigated by several research groups, e.g. 11, 29, 46] However, only little experience for the use of non predictable data availability has been made so far, e.g. ....

Anderson, D.P.: "Meta Scheduling for Distributed Continuous Media", Report No. UCB/CSD 90/599, Computer Science Division, University of California Berkeley, October 1990

....amount of related work targeted on focused problems. This work is of high interest for us, especially those addressing, synchronization in general, e.g. 18, 27, 28] and especially for distributed environments, e.g. 22, 23, 24, quality of service issues, e.g. 9, 19] and scheduling, e.g. [3, 6, 12]. Regarding the classification of synchronization techniques proposed in [8] our approach belongs to the local schemes since it does not implement network protocol based synchronization. Currently, it is not entirely saddled yet, however, it is similar to the one of the ACME system [2] However, ....

Anderson, D.P.: "Meta-Scheduling for Distributed Continuous Media", Report No. UCB/CSD 90/599, Computer Science Division, University of California Berkeley, October 1990

....However, the clients may wish to specify some ordering in this degradation. 13 based on a time stamping model while we plan to build related parameters channel groups on the basis of jitter controlled channels. The strands of Rangan and Vin [RanVin91] and the ropes of the Berkeley DASH project [Anderson90] are also similar to bundles. Campbell et al. describe an orchestration service for coordinating multiple related transport virtual circuits in [CaCoGaHu92] They say that the reason why they employ the term orchestration rather than synchronization is because cross stream relationships may ....

D. Anderson, "Meta-scheduling for Distributed Continuous Media", Technical Report UCB/CSD 90/599, University of California, Berkeley, CA (Oct 1990)

....bandwidth among the cooperating streams belonging to an application. This allocation is changed over time to match the dynamic nature of the streams. We use the M LBAP traffic characterization model to specify the characteristics of the streams. The M LBAP model is a variation of the LBAP model (Anderson 1993) that gives tighter characterization for the traffic of individual streams. The rest of this paper is organized as follows. In section 2, the M LBAP traffic characterization model is described. In section 3, delay bounds for streams sharing the same group reservation are derived analytically. A ....

....model (oe; ae) were developed. Central to the analysis is the concept of traffic constraint function b(t) b(t) is defined to be the maximum number of bits that can arrive during any interval of length t. For the (oe; ae) model, b(t) oe aet. The linear bounded arrival processes (LBAP) model (Anderson 1993), characterizes the traffic using three parameters (R, B, S) where R is the average rate in bits sec, B is the maximum burst size is packets, and S is the maximum packet size in bits. It can be easily shown that the LBAP model is simply a (oe; ae) model with oe = BS and ae = R. The LBAP model has ....

Anderson, D.P.(1993) Meta-scheduling for distributed continuous media.

....results from our telerobotics experiment. Section 5 focuses on implementation issues, and Section 6 concludes the paper. 2 Brief Overview of QoS Broker OMEGA Architecture To provide application to application guarantees, we need guaranteed services both in the network and at the end points [And93] We assume that our communication architecture resides on top of guaranteed network services, as many results illustrating methods to provide such services are now mature. We concentrate our research efforts on providing end to end QoS guarantees which rely on the end to end resource management. ....

D. P. Anderson. Meta-Scheduling for Distributed Continuous Media. ACM Transaction on Computer Systems, 11(3), August 1993.

....To support integrated continuous multimedia, resources such as the central processor, peripheral processors (audio, video) and communication network capacity must be allocated and scheduled together to provide the necessary data rate guarantees. This meta scheduling has been studied by Anderson [29]. 7 Conclusions This paper presents two studies conducted to validate Swift, a scalable distributed I O architecture that achieves high data rates by striping data across several storage agents and driving them concurrently. The prototype validates the concept of distributed disk striping in a ....

D. Anderson, "Meta-scheduling for distributed continuous media," Tech. Rep. UCB/CSD 90/599, University of California, Berkeley, Oct. 1990.

....approach. Each client enqueues received data on its respective queue, but the realtime scheduler determines which thread(s) to run next based on the information given by the realtime clients. One applicable realtime scheduling discipline is the Deadline Workahead Scheduling of Anderson [1990], where realtime processes are critical if they have unprocessed messages 12 A Fast, Flexible Network Interface Framework that are within a deadline based on the message s logical (not actual) arrival time. An enhanced EarlyDemuxClient, when invoked by the driver at interrupt time, can check to ....

Anderson, D.P. (1990) Meta-scheduling for distributed continuous media. Technical Report UCB/CSD 90/599, University of California at Berkeley EECS Department.

....system architectures along with corresponding algorithms. 1 Introduction An adaptive execution of multimedia presentations can be required for several reasons. For example, one reason is that although there exists promissing work on resource reservation for multimedia presentations, e.g. [2, 7, 13, 15, 10, 35, 24], it is likely that within the next decade most general purpose computing environments will not fully support end to end quality of service (QoS) guarantees. In order to also allow in those environments the execution of multimedia presentations 1 performance fluctuations must be handled. This ....

Anderson, D.P.: "Meta Scheduling for Distributed Continuous Media", Report No. UCB/CSD 90/599, Computer Science Division, University of California Berkeley, October 1990

....among the cooperating streams belonging to the same application. This allocation is changed over time to match the dynamic nature of the streams. We devised the M LBAP traffic characterization model to specify the characteristics of the streams. The MLBAP model is a variation of the LBAP model [3, 12] that gives tighter characterization for the traffic of individual streams. The QoSess control layer uses information about the traffic characteristics of the streams, their ranges of operation in terms of rate and delay bounds, their current relative priorities, and the total reserved resources, ....

....on a fluid traffic model (oe; ae) Central to the analysis is the concept of traffic constraint function b(t) b(t) is defined to be the maximum number of bits that can arrive during any interval of length t. For the (oe; ae) model, b(t) oe aet. The linear bounded arrival processes (LBAP) model [3, 12], characterizes the traffic using three parameters (R, B, S) where R is the average rate in bits sec, B is the maximum burst size is packets, and S is the maximum packet size in bits. It can be easily shown that the LBAP model is simply a (oe; ae) model with oe = BS and ae = R. The LBAP model has ....

D.P. Anderson, "Meta-scheduling for distributed continuous media", ACM Transactions on Computer Systems, 11(3), August 1993.

....for particular devices, for instance CPU cycles; network throughput, bit error rate, and delay jitter; disk bandwidth; resolution and colors of the display. In communicating with devices (for measuring or requesting quality) and admission testers a multimedia system needs to use these dimensions [2, 3, 34, 41, 42]. They are mostly meaningless to end users. 2.1.4 Mapping between QoS Levels Any sophisticated QoS management (with or without performance guarantees) needs a way to assess the quality a resource level QoS configuration provides. This assessment is necessary to compare and choose among several ....

....Proposed Solutions Protocols A variety of network protocols for real time transmission and bandwidth reservation have been developed. Guarantees can be provided only if all routers on a path through the network support a protocol, however. The session reservation protocol SRP in the DASH system [2, 3] processes resources in two phases. In the first phase, all nodes in the pipeline reserve the resources involved. If a maximum end to end delay is exceeded, the reservation fails. On the other hand, if the delay is smaller than a target delay, the reservations are relaxed in a second phase. The ....

[Article contains additional citation context not shown here]

D. P. Anderson. Meta-scheduling for distributed continuous media. Technical report CB/CSD 90/599, University of California, Berkeley, October 1990.

....is available showing whether the expressive power of these schemes meets the requirements in advanced multimedia applications. Low level synchronization of continuous data streams is also a service which will be offered to some extend by advanced networks and communication systems [AH91, And90] In addition to the synchronization of processing of the media data, concepts are needed to model the interaction with the user of the data, e.g. a user viewing a multimedia presentation. Interaction with the user may include standard functions like stop, start, continue, pause, but also ....

D. P. Anderson. Meta-scheduling for Distributed Continuous Media. Technical Report No. UCB/CSD 90/599. EECS, University of California at Berkeley, Berkeley, CA, USA, 1990.

....error handling and for maintaining proper timing relationships within and between CM streams. 6 In addition, the elimination of synchronous interactions between the client and the service may simplify data transmission for many classes of clients. For example, data from a CM file server, e.g. [And90] or a hardware compression module could be transferred directly to the transport service without an intervening process. At the receiving side, the CM transport service can write data directly into a buffer shared with a multimedia display server that is managing input and output devices and ....

D. Anderson, "Meta-Scheduling for Distributed Continuous Media", UC Berkeley, EECS Dept., Technical Report No. UCB/CSD 90/599, (October 1990).

....quality of service guarantees between the human user and the remote devices. To provide such guarantees, networked applications need to either directly or indirectly control network resources (e.g. in switches) and end point resources such as memory, CPU and I O device access[And93]. Resource control must be integrated in a robust admission and scheduling process. We assume that such admission and scheduling exist for network resources such as bandwidth and delay, and focus on an admission and scheduling process for the end points. By integrating the endpoint resources, we ....

....but not sufficient. For example, the fact that data has arrived at a network adapter on time does not imply that an application receives and processes the data in a timely manner. Real time communication must be integrated with real time computing support and applications at network end points [And93]. Anderson provided a uniform theoretical framework and model, called Metascheduler, for resource reservation to achieve application to application guarantees [And93] In practice, however, the applications and real time computing are still not fully integrated with the real time communication. ....

[Article contains additional citation context not shown here]

D. P. Anderson. Meta-Scheduling for Distributed Continuous Media. ACM Transaction on Computer Systems, 11(3), August 1993.

....system support in order to be extended to telemedicine and other advanced distributed virtual reality applications. Therefore, the primary goal of an underlying Networked Multimedia System (NMS) for these applications is to provide Quality of Service (QoS) guarantees in end to end fashion [And93] (Figure 1) Protocols) Communication (Operating System System Level Protocols) Communication (Operating System System Level Multimedia PC Multimedia Workstation END TO END QoS GUARANTEES System Architecture Multimedia Workstation Multimedia Workstation System Architecture Gateway FDDI Ring ATM ....

....into real memory to avoid page faults. Page faults seriously affect the real time performance. 5 Distributed Multimedia Protocols In a distributed environment, a setup mechanism that is integrated across all levels of the NMS is necessary to achieve guaranteed services in an end to end fashion [And93]. The setup mechanism integrates negotiation protocols, admission and translation services (to accommodate heterogeneity in quality representation at individual levels) media encoding of various multimedia devices, and schedulability of timing constraints. These services will be utilized by ....

D. P. Anderson. Meta-Scheduling for Distributed Continuous Media. ACM Transaction on Computer Systems, 11(3), August 1993.

....(e.g. workstations are equipped with a rich set of multimedia devices, and support multiprocessing and multiple users. This requires balancing resources among the application, network, and operating system within the endpoints as well as balancing resources between end points and the network [And93]. As a part of the OMEGA system, we designed and implemented a resource management entity, called the QoS Broker, which provides local and global resource management as shown in Figure 5. Network Resources Local Resources QoS Broker (BUYER) Application Subsystem Resources QoS Broker (SELLER) ....

....(RTNP) This mapping is done within the admission services. 3.3 Admission Admission control is an essential element to achieve guaranteed services. For distributed multimedia communications systems, each resource along the path(s) between source(s) and sink(s) must monitor its availability [And93]. In our OMEGA architecture, the control and monitor of end point resources is done by the QoS Broker, which performs admission control at both layers of the OMEGA system (Figure 2) For the admission and scheduling of RTAP and RTNP protocol data units (PDUs) we assume that (1) RTAP and RTNP ....

D. P. Anderson. Meta-Scheduling for Distributed Continuous Media. ACM Transaction on Computer Systems, 11(3), August 1993.

....thus the concepts and techniques of CMFS apply to each class. CMFS is meant to serve as part of a distributed system that handles integrated audio and video. End to end performance guarantees cannot, of course, be achieved by disk scheduling alone. CMFS conforms to the meta scheduling model [2], which provides a mechanism for making such guarantees. The role of CMFS in this larger context is discussed in Section 3. 2. CLIENT INTERFACE CMFS clients access real time files in sessions. Each session has a FIFO buffer for data transfer between the client and CMFS. For concreteness, assume ....

....handled by many shared hardware resources : disk, CPU, network, bus, memory, etc. To provide applications with deterministic end to end performance, we must integrate the scheduling of all these resources. CMFS is designed to serve as part of such a meta scheduling scheme, the CM resource model [2]. In this scheme, each resource can be reserved in sessions with fixed workload and delay bounds. The parameterization of workload and delay lets resources work ahead on real time streams so they can response quickly to non real time workload. A CMFS session s cushion parameter Y # allows ....

D. P. Anderson, "Meta-Scheduling for Distributed Continuous Media", UC Berkeley, EECS Dept., Technical Report No. UCB/CSD 90/599, Oct. 1990. 25

....process library. I O is done using UNIX asynchronous I O. The server handles telephone quality (64 Kbps) audio I O and video output, both compressed and uncompressed. 2.2. Deadline Workahead Scheduling The Deadline Workahead Scheduling (DWS) CPU scheduling policy is designed for integrated CM [2]. In the DWS model, a process that handles CM data is called a real time process. There are two classes of non real time processes: interactive (for which fast response time is important) and background. A real time process handles a sequence of messages each with a logical arrival time l (m ) ....

....variant, each new LWP specifies its workload (message rate and CPU time per message) The KLS conducts a schedulability test to determine whether the workload can be accommodated and if so, with what logical delay bound. This test involves a simulation under worst case load, and is described in [2]. In the soft variant, no such screening is done, and it is possible for the system to fall behind schedule. SLS is not restricted to deadline workahead scheduling; it can be adapted to other policies, such as static priorities or usage based timesharing policies. The policy dictates the contents ....

D. P. Anderson, "Meta-Scheduling for Distributed Continuous Media", UC Berkeley, EECS Dept., Technical Report No. UCB/CSD 90/599, Oct. 1990.

No context found.

D. Anderson, Meta-Scheduling for Distributed Continuous Media, Technical Report CSD-90-599, Computer Science Division, EECS Department, University of California at Berkeley, Berkeley, CA, USA, Oct. 1990.

First 50 documents

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