Output process of variable bit-rate flows in on-chip networks based on aggregate scheduling

Jafari, Fahimeh; Jantsch, Axel; Lu, Zhonghai

doi:10.1109/iccd.2011.6081442

Cited by 3 publications

(7 citation statements)

References 4 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Since the required theorems for calculating performance metrics of VBR traffic transmitted in the FIFO order and scheduled as aggregate have not been represented so far, we [Jafari et al 2011[Jafari et al , 2012 have defined and proved them based on network calculus [Chang 2000;Le Boudec et al 2004]. In Jafari et al [2011], we proposed and proved the required theorem for deriving the output characterization of VBR traffic under the defined system model to have exact vision about output metrics used for obtaining performance bounds. In Jafari et al [2012], the required theorems for computing end-to-end ESC and end-to-end delay bound are defined and proved.…”

Section: Introductionmentioning

confidence: 97%

See 1 more Smart Citation

Least Upper Delay Bound for VBR Flows in Networks-on-Chip with Virtual Channels

Jafari

Jantsch

2015

ACM Trans. Des. Autom. Electron. Syst.

Self Cite

View full text Add to dashboard Cite

Real-time applications such as multimedia and gaming require stringent performance guarantees, usually enforced by a tight upper bound on the maximum end-to-end delay. For FIFO multiplexed on-chip packet switched networks we consider worst-case delay bounds for Variable Bit-Rate (VBR) flows with aggregate scheduling, which schedules multiple flows as an aggregate flow. VBR Flows are characterized by a maximum transfer size (L), peak rate ( p), burstiness (σ ), and average sustainable rate (ρ). Based on network calculus, we present and prove theorems to derive per-flow end-to-end Equivalent Service Curves (ESC), which are in turn used for computing Least Upper Delay Bounds (LUDBs) of individual flows. In a realistic case study we find that the end-to-end delay bound is up to 46.9% more accurate than the case without considering the traffic peak behavior. Likewise, results also show similar improvements for synthetic traffic patterns. The proposed methodology is implemented in C++ and has low run-time complexity, enabling quick evaluation for large and complex SoCs.

show abstract

Section: Introductionmentioning

confidence: 97%

“…The method presented in Jafari et al [2012] only considers direct contentions of a tagged flow. In this article, we use the proposed theorems [Jafari et al 2011[Jafari et al , 2012 to present a formal approach for performance analysis modeling both direct and indirect contentions.…”

Section: Introductionmentioning

confidence: 99%

Least Upper Delay Bound for VBR Flows in Networks-on-Chip with Virtual Channels

Jafari

Jantsch

2015

ACM Trans. Des. Autom. Electron. Syst.

Self Cite

View full text Add to dashboard Cite

show abstract

“…1) Delay Optimization: As mentioned before, decision variables in the proposed optimization problems are the weights on shared channels. Due to shared channels in VOPD application, 20 weights are formulated in the optimizations as a weight vector W defined as follows: W = w (6,3,0,4) , w (10,2,0,0) , w (14,0,0,2) , w (13,3,0,2) , w (12,0,0,2) , w (9,4,0,0) , w (4,3,0,4) , w (4,0,0,2) , w (8,2,0,0) , w (8,4,0,2) , w (6,2,0,4) , w (10,4,0,0) , w (14,3,0,2) , w (13,1,0,2) , w (12,3,0,2) , w (9,3,0,0) , w (4,0,0,4) , w (4,4,0,2) , w (8,…”

Section: A Vopd Applicationmentioning

confidence: 99%

“…Due to shared channels in the MMS application, weight vector W consisting of 37 weights is defined as below: W = w (16,1,0,0) , w (15,3,0,2) , w (15,0,0,2) , w (5,3,0,0) , w (14,4,0,2) , w (19,1,0,2) , w (18,4,0,2) , w (18,3,0,2) , w (13,2,0,4) , w (23,3,0,0) , w (14,1,0,4) , w (10,2,0,4) , w (1,0,0,4) , w (1,3,0,…”

Section: B Mutlimedia Applicationmentioning

confidence: 99%

“…Therefore, the first important consideration is to derive the LUDB for a given communication flow. To this end, based on the Network Calculus theory [1], we have presented and proved required theorems in [2] [3]. We then presented a methodology [4] to consider resource sharing scenarios and also derive end-toend Equivalent Service Curve (ESC) and LUDB by applying the proposed theorems.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Weighted Round Robin Configuration for Worst-Case Delay Optimization in Network-on-Chip

Jafari

Jantsch

2016

IEEE Trans. VLSI Syst.

Self Cite

View full text Add to dashboard Cite

Abstract-We propose an approach for computing the endto-end delay bound of individual variable bit-rate flows in a FIFO multiplexer with aggregate scheduling under Weighted Round Robin (WRR) policy. To this end, we use network calculus to derive per-flow end-to-end equivalent service curves employed for computing Least Upper Delay Bounds (LUDBs) of individual flows. Since real time applications are going to meet guaranteed services with lower delay bounds, we optimize weights in WRR policy to minimize LUDBs while satisfying performance constraints. We formulate two constrained delay optimization problems, namely, Minimize-Delay and Multiobjective optimization. Multi-objective optimization has both total delay bounds and their variance as minimization objectives. The proposed optimizations are solved using a genetic algorithm. A Video Object Plane Decoder (VOPD) case study exhibits 15.4% reduction of total worst-case delays and 40.3% reduction on the variance of delays when compared with round robin policy. The optimization algorithm has low run-time complexity, enabling quick exploration of large design spaces. We conclude that an appropriate weight allocation can be a valuable instrument for delay optimization in on-chip network designs.

show abstract