• Corpus ID: 102350728

ReNets: Toward Statically Optimal Self-Adjusting Networks

  title={ReNets: Toward Statically Optimal Self-Adjusting Networks},
  author={C. Avin and Stefan Schmid},
This paper studies the design of self-adjusting networks whose topology dynamically adapts to the workload, in an online and demand-aware manner. This problem is motivated by emerging optical technologies which allow to reconfigure the datacenter topology at runtime. Our main contribution is ReNet, a self-adjusting network which maintains a balance between the benefits and costs of reconfigurations. In particular, we show that ReNets are statically optimal for arbitrary sparse communication… 

Figures from this paper

Working Set Theorems for Routing in Self-Adjusting Skip List Networks

This paper presents SANs which provide, for the first time, provable working set guarantees: the routing cost between node pairs is proportional to how recently these nodes communicated last time, and presents deterministic SAN algorithms.

Online Dynamic B-Matching∗ With Applications to Reconfigurable Datacenter Networks

This paper initiates the study of online algorithms for the maximum weight b-matching problem, a generalization of maximum weight matching where each node has at most b ≥ 1 adjacent matching edges, and proves that this is asymptotically optimal.

Online Dynamic B-Matching

This paper introduces the study of online algorithms for the maximum weight b-matching problem, a generalization of maximum weight matching where each node has at most b≥1 adjacent matching edges, and develops an online algorithm which is O(b)- competitive and proves that this is asymptotically optimal.

Scheduling Opportunistic Links in Two-Tiered Reconfigurable Datacenters

A stable-matching algorithm is presented which is O(ε^-2 )-competitive against an optimal offline algorithm, in a resource augmentation model: the online algorithm runs 2+ε times faster.

On the Complexity of Non-Segregated Routing in Reconfigurable Data Center Architectures

It is shown that in certain specific scenarios, optimal data center topologies with nonsegregated routing policies can be computed in polynomial-time, and several important variants to be NP-hard.

SplitCast: Optimizing Multicast Flows in Reconfigurable Datacenter Networks

SplitCast is presented, a preemptive multicast scheduling approach that fully exploits emerging physical-layer multicast capabilities to reduce flow times and leverages flexibilities by supporting splittable multicast so that a transfer can already be delivered to just a subset of receivers when the circuit capacity is insufficient.

Survey of Reconfigurable Data Center Networks: Enablers, Algorithms, Complexity

This paper provides an overview of the algorithmic problems introduced by this technology, and surveys rst solutions.

Optimizing multicast flows in high-bandwidth reconfigurable datacenter networks

On the Complexity of Traffic Traces and Implications

Using a case study in the context of datacenters, it is shown that insights into the structure of packet traces can lead to improved demand-aware network designs: datacenter topologies that are optimized for specific traffic patterns.

On the Complexity of Traffic Traces and Implications

  • C. AvinM. GhobadiChen GrinerS. Schmid
  • Computer Science
    Abstracts of the 2020 SIGMETRICS/Performance Joint International Conference on Measurement and Modeling of Computer Systems
  • 2020
This paper introduces the notion of trace complexity which approximates the entropy rate of a packet trace, and proposes a traffic generator model able to produce a synthetic trace that matches the complexity levels of its corresponding real-world trace.



SplayNet: Towards Locally Self-Adjusting Networks

This paper introduces a simple model which captures the fundamental tradeoff between the benefits and costs of self-adjusting networks, and presents the SplayNet algorithm and formally analyze its performance, and proves its optimality in specific case studies.

Demand-Aware Network Design with Minimal Congestion and Route Lengths

This paper presents the first bounded-degree, demand-aware network, ct-DAN, which minimizes both congestion and route lengths and is provably (asymptotically) optimal in each dimension individually.

Demand-aware network designs of bounded degree

A general lower bound based on the entropy of the communication pattern of a bounded-degree network is derived, and asymptotically optimal demand-aware network design algorithms for important distribution families, such as sparse distributions and distributions of locally bounded doubling dimensions are presented.

Jellyfish: Networking Data Centers Randomly

Jellyfish is a high-capacity network interconnect which, by adopting a random graph topology, yields itself naturally to incremental expansion, and is more cost-efficient than a fat-tree.

Fat-FREE Topologies

It is taken that alongside the development of reconfigurable topologies, the community should also invest in developing superior traffic engineering over static networks other than fat-trees as an alternate path forward, and a rigorous quantification of the difference between the power of two techniques for handling dynamic, unpredictable traffic with limited network resources.

Achieving high utilization with software-driven WAN

A novel technique is developed that leverages a small amount of scratch capacity on links to apply updates in a provably congestion-free manner, without making any assumptions about the order and timing of updates at individual switches.

Beyond fat-trees without antennae, mirrors, and disco-balls

The results substantially lower the barriers for improving upon today's data centers by showing that a static, cabling-friendly topology built using commodity equipment yields superior performance when combined with well-understood routing methods.

Proteus: a topology malleable data center network

Proteus is proposed, an all-optical architecture targeting unprecedented topology-flexibility, lower complexity and higher energy efficiency, that is, an oversubscribed network architectures that can adapt the topology to meet traffic demands.

Flyways To De-Congest Data Center Networks

A study of application demands from a production datacenter of servers shows that except for a few outliers, application demands can be generally met by a network that is slightly oversubscribed, and advises a hybrid architecture to tackle the hotspots that remain.

Joint Static and Dynamic Traffic Scheduling in Data Center Networks

The observation that the traffic in a data center is a mixture of relatively static and rapidly fluctuating components is used to develop a combined scheduler for both these components using a generalization of the load-balanced scheduler.