Hierarchical packet fair queueing algorithms

Bennett, Jon C. R.; Zhang, Hui

doi:10.1109/90.649568

Cited by 369 publications

(204 citation statements)

References 22 publications

Supporting

Mentioning

201

Contrasting

Unclassified

Order By: Relevance

“…While most of the previous research directed at providing better services in packet switching networks have focused on providing guaranteed services or protection for each individual flow, several recent works [11,39,92] have argued that it is also important to support hierarchical link-sharing service.…”

Section: Link Sharingmentioning

confidence: 99%

See 1 more Smart Citation

Stateless Core: A Scalable Approach for Quality of Service in the Internet

Stoica¹,

Zhang²

2004

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Today's Internet provides one simple service: best effort datagram delivery. This minimalist service allows the Internet to be stateless, that is, routers do not need to maintain any fine grained information about traffic. As a result of this stateless architecture, the Internet is both highly scalable and robust. However, as the Internet evolves into a global commercial infrastructure that is expected to support a plethora of new applications such as IP telephony, interactive TV, and e-commerce, the existing best effort service will no longer be sufficient. In consequence, there is an urgent need to provide more powerful services such as guaranteed services, differentiated services, and flow protection.Over the past decade, there has been intense research toward achieving this goal. Two classes of solutions have been proposed: those maintaining the stateless property of the original Internet (e.g., Differentiated Services), and those requiring a new stateful architecture (e.g., Integrated Services). While stateful solutions can provide more powerful and flexible services such as per flow bandwidth and delay guarantees, they are less scalable than stateless solutions. In particular, stateful solutions require each router to maintain and manage per flow state on the control path, and to perform per flow classification, scheduling, and buffer management on the data path. Since today's routers can handle millions of active flows, it is difficult, if not impossible, to implement such solutions in a scalable fashion. On the other hand, while stateless solutions are much more scalable, they offer weaker services.The key contribution of this dissertation is to bridge this long-standing gap between stateless and stateful solutions in packet switched networks such as the Internet. Our thesis is that "it is actually possible to provide services as powerful and as flexible as the ones implemented by a stateful network using a stateless network". To prove this thesis, we propose a novel technique called Dynamic Packet State (DPS). The key idea behind DPS is that, instead of having routers maintain per flow state, packets carry the state. In this way, routers are still able to process packets on a per flow basis, despite the fact that they do not maintain any per flow state. Based on DPS, we develop a network architecture called Stateless Core (SCORE) in which core routers do not maintain any per flow state. Yet, using DPS to coordinate actions of edge and core routers along the path traversed by a flow allows us to design distributed algorithms that emulate the behavior of a broad class of stateful networks in SCORE networks.In this dissertation we describe complete solutions including architectures, algorithms and implementations which address three of the most important problems in today's Internet: providing guaranteed services, differentiated services, and flow protection. Compared to existing solutions, our solutions eliminate the most complex operations on both the data and control paths in the network core, i...

show abstract

Section: Link Sharingmentioning

confidence: 99%

“…Due to the service complexity, it should come as no surprise that all current solutions require routers to maintain per class state [11,39,92]. A natural research direction would be to implement the link-sharing service in a SCORE network.…”

Section: Link Sharingmentioning

confidence: 99%

Stateless Core: A Scalable Approach for Quality of Service in the Internet

Stoica¹,

Zhang²

2004

Lecture Notes in Computer Science

View full text Add to dashboard Cite

show abstract

“…In effect, items of higher access probability are assigned to faster disks than the items of low access probability, which are assigned to slow revolving disks. Later, Jain and Werth [18] proved that the optimal expected access time results when the instances of each item to be pushed are equally spaced, while Bennett and Zhang [9] determined which packet from many input queue should be transmitted next in the output channel so that the channel is used in a fair way. The fact that the push broadcast scheduling problem was related to Packet Fair Queuing was brought up by Vaidya and Hameed in [12,14], who also studied scheduling for multiple broadcast channels and the impact of the transmission error on scheduling.…”

Section: State Of the Art: Past Related Workmentioning

confidence: 99%

Push Less and Pull the Current Highest Demanded Data Item to Decrease the Waiting Time in Asymmetric Communication Environments

Pinotti

Saxena

2002

Lecture Notes in Computer Science

View full text Add to dashboard Cite

In these days, the world experiences an unprecedented demand for data and data services, driven mainly by the popularity of the Web services and by the evolution of the Internet towards an information super-highway. We introduce a hybrid scheduling that effectively combines broadcasting for very popular data (push data) and dissemination upon-request for less popular data (pull data) in asymmetric communication environments. In our solution, the server continuously broadcasts one push item and disseminates one pull item. The clients send their requests to the server, which queues-up them for the pull items. At any instant of time, the item to be broadcast is designated applying a purepush scheduling, while the item to be pulled is the one stored in the pull-queue, which has accumulated, so far, the highest number of pending requests. The value of the average expected waiting time spent by a client in the hybrid system, denoted by T exp-hyb , is evaluated analytically, and the cut-off point between push and pull items is chosen in such a way that T exp-hyb is minimized. We find out that by doing so we can drop the cut off point to a value, which is much less than the total number of items present in the system, improving upon the average waiting time spent by a client in a pure push system and also on that spent in some of the hybrid systems already proposed in literature. IntroductionDay by day the ability to interconnect computers through cable, satellite and wireless networks is increasing and proportionately to this is also increasing a new application based on data dissemination. This application focuses on delivering data to a large population of clients. Often in dissemination-based systems, there exists communications' asymmetry. This asymmetry may arise due to several factors like:The downstream communication capacity (bandwidth from server to client) may be much greater than the upstream communication capacity (bandwidth from client to server); In Information Retrieval Applications, the clients make requests to the server through small request messages that result in the transfer of much larger objects; Systems with small number of servers an d large number of clients also result in such asymmetry. Basically, there are two approaches to spread data items in such systems: the Push-based data scheduling, and the pull-based data scheduling. A system where the client simply grabs the data being broadcast without making any requests is an example of push-based system. In such systems, the clients continuously monitor the broadcast process and retrieve the data items they require. The server, on the other hand, broadcasts data items on scheduled time no matter whether the particular item is being required at that time or not. On the contrary, pull-based systems are on demand traditional client server systems where clients and server have a request/response style of relationship. In such systems, the clients initiate the data transfer by sending requests and the server then makes a schedule to satisfy ...

show abstract

“…Hierarchical resource sharing or dynamic storage allocation also finds its analogy in link-sharing in the network transmission context [54,82]. A content owner may have different classes of objects in its corpus, and wishes to assign different QoS levels for the different classes.…”

Section: Hierarchical Resource Sharingmentioning

confidence: 99%

Distributed network storage service with quality-of-service guarantees

Chuang

Sirbu

2000

Journal of Network and Computer Applications

View full text Add to dashboard Cite

This paper envisions a distributed network storage service with Quality-ofService (QoS) guarantees, and describes its architecture and key mechanisms. When fully realized, this service architecture would be able to support, in one integrated framework, network storage services ranging from best-effort caching to replication with performance guarantees. Content owners could, through the use of standardized protocols, reserve network storage resources to satisfy their application-specific performance requirements. They would be able to specify either the number and/or placement of the replicas, or higher-level performance goals based on access latency, bandwidth usage or data availability. The network storage provider would then optimally allocate storage resources to meet the service commitments, using leftover capacity for best-effort caching. Content consumers would then retrieve the nearest copy of the data object, be it from a replica, cache, or the original source, in a completely transparent manner. Furthermore, a distributed network storage infrastructure should be integrated with the existing transmission-based QoS framework so that applications can select the optimal combination of storage and transmission resources to satisfy their performance requirements.This work identifies and discusses key research areas and problems that need to be tackled, including those in service specification, resource mapping, admission control, resource reservation and real-time storage management. The paper establishes a QoS framework upon which community discussion on this vision can proceed.

show abstract

Hierarchical packet fair queueing algorithms

Cited by 369 publications

References 22 publications

Stateless Core: A Scalable Approach for Quality of Service in the Internet

Stateless Core: A Scalable Approach for Quality of Service in the Internet

Push Less and Pull the Current Highest Demanded Data Item to Decrease the Waiting Time in Asymmetric Communication Environments

Distributed network storage service with quality-of-service guarantees

Contact Info

Product

Resources

About