Fair queuing for aggregated multiple links

Blanquer, Josep M.; Özden, Banu

doi:10.1145/383059.383074

Cited by 54 publications

(30 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…4 With the same assumptions as in Theorem 4.1, the bounds in Equation (6) and Equation (7) are tight. More precisely, we can always define a succession of packets and a series of scheduling behaviors of the chain of GR nodes such that the burstiness of the flow at the output of the M -th node achieves the bound in Equation (7), and that at least one packet from the given flow experiences an end-to-end delay equal to the bound in Equation (6).…”

Section: The Delay Bound In the Non-fifo Case Is Tightmentioning

confidence: 96%

“…Thus the GR node model of a differentiated services router cannot always be assumed to be FIFO per flow. More generally, the GR class encompasses a great variety of algorithms, which are not necessarily FIFO per flow [4]. In this paper we use the term "FIFO" to indicate a GR node that is FIFO per flow (since the definition of GR node is relative to the treatment it gives to a flow viewed as a single entity).…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

“Pay bursts only once” does not hold for non-FIFO guaranteed rate nodes

Rizzo

Boudec

2005

Performance Evaluation

View full text Add to dashboard Cite

We consider end-to-end delay bounds in a network of Guaranteed Rate (GR) nodes. We demonstrate that, contrary to what is generally believed, the existing end-to-end delay bounds apply only to GR nodes that are FIFO per flow. We show this by exhibiting a counter-example.Then we show that the proof of the existing bounds has a subtle, but important, dependency on the FIFO assumption, which was never noticed before. Finally, we give a tight delay bound that is valid in the non-FIFO case; it is noticeably higher that the existing one. In particular, the phenomenon known as "pay bursts only once" does not apply to non-FIFO nodes. These findings are important in the context of differentiated services. Indeed the existing bounds have been applied to cases where a flow (in the sense of the GR definition) is an aggregate of end-user microflows, and it is not generally true that a router is FIFO per aggregate; thus the GR node model of a differentiated services router cannot always be assumed to be FIFO per flow.

show abstract

Section: The Delay Bound In the Non-fifo Case Is Tightmentioning

confidence: 96%

Section: Introductionmentioning

confidence: 99%

“Pay bursts only once” does not hold for non-FIFO guaranteed rate nodes

Rizzo

Boudec

2005

Performance Evaluation

View full text Add to dashboard Cite

show abstract

“…We assumed the network to have topology as shown in Fig. 6 which is similar to standard Cisco campus network 6 . It is 3-Tier architecture with 3136 servers and 260 routers.…”

Section: Trace-driven Simulationmentioning

confidence: 99%

“…Overall, fair bandwidth allocation is a well studied topic in computer networks [6]. Researchers proposed solutions typically emulating max-min allocations to tenants treating bandwidth as the same interchangeable resource.…”

Section: Related Workmentioning

confidence: 99%

STEM+: Allocating bandwidth fairly to tasks

Nikolaevskiy

Lukyanenko

Gurtov

2015

2015 IEEE Conference on Computer Communications Workshops (INFOCOM WKSHPS)

View full text Add to dashboard Cite

Fair sharing of bandwidth among tenants in datacenters is important to guarantee prompt execution while providing isolation between different jobs. Existing bandwidth allocation methods lack a concept of a task reflecting the dependency between allocations on links. Moreover, existing approaches do not consider the tenants to be smart individuals and lack understanding of a threat that strategic players can produce. In this work we introduce a Strategy-proof Task-Enforcement Mechanism (STEM) which is the only strategy-proof mechanism for datacenter allocation. It seamlessly utilizes task-aware models. While tenants are able to improve their allocations by relocating demands among links, it also improves the global allocation resulting into a strong Nash equilibrium among tenants. This is in contrast to pricing or Competitive Equilibrium from Equal Incomes (CEEI) which permits tenants to inflate their demands and in some cases loosing sharing-incentives. We extend STEM with STEM+ -a work-conserving allocation mechanism.

show abstract

“…Finally, if the scheduler providing PRSG to an aggregate is more complicated than a FIFO, the scheduler itself may cause packet reordering. An example of such a scheduler is described in [16].…”

Section: Introductionmentioning

confidence: 99%

Packet scale rate guarantee for non-fifo nodes

Boudec

Charny

2003

IEEE/ACM Trans. Networking

View full text Add to dashboard Cite

Abstract-Packet Scale Rate Guarantee (PSRG) is a generic node model which underlies the definition of Expedited Forwarding (EF) proposed in the context of Internet Differentiated Services. For the case of FIFO nodes, PSRG is equivalent to the well-understood concept of adaptive service curve. However, in practice, many devices do not necessarily preserve the FIFO property, and therefore known FIFO results do not hold. This paper analyzes the properties of PSRG in the absence of FIFO assumption. Our analysis is based on a novel characterization of PSRG which avoids the use of virtual finish times; it is obtained by min-max algebra. We use it to show that delay bounds previously obtained for the FIFO case are still valid; in contrast, we find that this is not true for the characterization of the concatenation of two nodes.

show abstract

Fair queuing for aggregated multiple links

Cited by 54 publications

References 9 publications

“Pay bursts only once” does not hold for non-FIFO guaranteed rate nodes

“Pay bursts only once” does not hold for non-FIFO guaranteed rate nodes

STEM+: Allocating bandwidth fairly to tasks

Packet scale rate guarantee for non-fifo nodes

Contact Info

Product

Resources

About