| 1. |
- Hernandez Benet, Cristian, et al.
(författare)
-
Providing In-network Support to Coflow Scheduling
- 2021
-
Ingår i: Proceedings of the 2021 IEEE Conference on Network Softwarization: Accelerating Network Softwarization in the Cognitive Age, NetSoft 2021. - : IEEE. - 9781665405225 ; , s. 235-243
-
Konferensbidrag (refereegranskat)abstract
- Many emerging distributed applications, including big data analytics, generate a number of flows that concurrently transport data across data center networks. To improve their performance, it is required to account for the behavior of a collection of flows, i.e., coflows, rather than individual. State-of-the-art solutions allow for a near-optimal completion time by continuously reordering the unfinished coflows at the end-host, using network priorities. This paper shows that dynamically changing flow priorities at the end host, without taking into account in-flight packets, can cause high-degrees of packet re-ordering, thus imposing pressure on the congestion control and potentially harming network performance in the presence of switches with shallow buffers. We present pCoflow, a new solution that integrates end-host based coflow ordering with in-network scheduling based on packet history. Our evaluation shows that pCoflow improves in CCT upon state-of-the-art solutions by up to 34% for varying load.
|
|
| 2. |
- Pascual, D., et al.
(författare)
-
Accounting for Winter Warming Events in the Ecosystem Model LPJ-GUESS : Evaluation and Outlook
- 2024
-
Ingår i: Journal of Geophysical Research: Biogeosciences. - 2169-8953. ; 129:3
-
Tidskriftsartikel (refereegranskat)abstract
- Winter warming events (WWEs) are short-lasting events of unusually warm weather, occasionally combined with rainfall, which can cause severe ecosystem impacts by altering ground temperatures and water fluxes. Despite their importance, how large-scale ecosystem models perform in depicting the impacts of WWEs remain largely unknown. The frequency and intensity of WWEs will likely increase further in the future, making it necessary to understand their potential impacts on high-latitude ecosystems. In this study, we evaluated the ability of the dynamic ecosystem model Lund-Potsdam-Jena General Ecosystem Simulator (LPJ-GUESS) to represent the responses of subarctic ecosystems to future WWEs, and identified model gaps hindering more accurate estimates of these responses. In response to WWEs, the model simulated substantial ground cooling (up to 2°C in winter) due to reduced snow depth (insulation), with rain on snow (ROS) exerting a marginal influence on the ground temperature responses: these modeled responses are in apparent contradiction with the strong ground warming effect of ROS reported in most observational studies. The simulated ground cooling led to changes in biogeochemical fluxes that were substantial and often comparable in magnitude (but often opposite in direction) to those from altered winter climatologies. The mismatch between the modeled and the observed ground temperature responses to WWEs highlights LPJ-GUESS's current limitations in realistically simulating some of the effects of WWEs. These limitations likely stem from the (a) absence of a surface energy balance, (b) lack of snow-vegetation interactions, (c) daily time-step, and (d) simplistic water retention scheme in LPJ-GUESS.
|
|
| 3. |
- Pongracz, Alexandra, et al.
(författare)
-
Model simulations of arctic biogeochemistry and permafrost extent are highly sensitive to the implemented snow scheme in LPJ-GUESS
- 2021
-
Ingår i: Biogeosciences. - : Copernicus GmbH. - 1726-4170 .- 1726-4189. ; 18:20, s. 5767-5787
-
Tidskriftsartikel (refereegranskat)abstract
- The Arctic is warming rapidly, especially in winter, which is causing large-scale reductions in snow cover. Snow is one of the main controls on soil thermodynamics, and changes in its thickness and extent affect both permafrost thaw and soil biogeochemistry. Since soil respiration during the cold season potentially offsets carbon uptake during the growing season, it is essential to achieve a realistic simulation of the effect of snow cover on soil conditions to more accurately project the direction of arctic carbon-climate feedbacks under continued winter warming. The Lund-Potsdam-Jena General Ecosystem Simulator (LPJ-GUESS) dynamic vegetation model has used - up until now - a single layer snow scheme, which underestimated the insulation effect of snow, leading to a cold bias in soil temperature. To address this shortcoming, we developed and integrated a dynamic, multi-layer snow scheme in LPJ-GUESS. The new snow scheme performs well in simulating the insulation of snow at hundreds of locations across Russia compared to observations. We show that improving this single physical factor enhanced simulations of permafrost extent compared to an advanced permafrost product, where the overestimation of permafrost cover decreased from 10% to 5% using the new snow scheme. Besides soil thermodynamics, the new snow scheme resulted in a doubled winter respiration and an overall higher vegetation carbon content. This study highlights the importance of a correct representation of snow in ecosystem models to project biogeochemical processes that govern climate feedbacks. The new dynamic snow scheme is an essential improvement in the simulation of cold season processes, which reduces the uncertainty of model projections. These developments contribute to a more realistic simulation of arctic carbon-climate feedbacks.
|
|
| 4. |
|
|
