Towards predictable datacenter networks

@article{Ballani2011TowardsPD,
  title={Towards predictable datacenter networks},
  author={Hitesh Ballani and Paolo Costa and Thomas Karagiannis and Antony Ian Taylor Rowstron},
  journal={Proceedings of the ACM SIGCOMM 2011 conference},
  year={2011}
}
The shared nature of the network in today's multi-tenant datacenters implies that network performance for tenants can vary significantly. This applies to both production datacenters and cloud environments. Network performance variability hurts application performance which makes tenant costs unpredictable and causes provider revenue loss. Motivated by these factors, this paper makes the case for extending the tenant-provider interface to explicitly account for the network. We argue this can be… Expand
On the impact of the network hypervisor on virtual network performance
TLDR
It is shown that cloud application performance may appear unpredictable if the network hypervisor is not accounted for, and hypervisors should be included in performance models, and their performance benchmarked and compared similarly to other crucial software components such as the SDN controller. Expand
Chatty Tenants and the Cloud Network Sharing Problem
TLDR
This work argues for network allocations to be dictated by the least-paying of communication partners, and shows that minimum bandwidth guarantees, apart from helping tenants achieve predictable performance, also improve overall datacenter throughput. Expand
omniCluster-Virtual Network Allocation in Data Centers using Software-Defined Networking
The creation of data centers allowed global access to huge computational resources, previously only available to large companies or governments. By renting the desired computational power, smallExpand
The only constant is change: incorporating time-varying network reservations in data centers
In multi-tenant datacenters, jobs of different tenants compete for the shared datacenter network and can suffer poor performance and high cost from varying, unpredictable network performance.Expand
Managing Network Reservation for Tenants in Oversubscribed Clouds
TLDR
This work presents a novel network reservation mechanism that can admit time-varying tenant requests and can fairly distribute any degradation in the NSR among the tenants in presence of network over subscription. Expand
Joint Virtual Machine Placement and Path Selection in a Virtualized Datacenter Environment
TLDR
This work designs a modified greedy approach that uses virtual machine placement and path selection to solve the joint resource allocation problem and designs a rigorous simulation setup to derive a definitive set of guidelines for the datacenter designer implementing the proposed SLA. Expand
Providing guaranteed network performance across tenants: Advances, challenges and opportunities
The varied network performance in the cloud hurts application performance. This increases the tenant's cost and becomes the key hindrance to cloud adoption. It is because virtual machines (VMs)Expand
Allocating Bandwidth in Datacenter Networks: A Survey
  • L. Chen, Baochun Li, Bo Li
  • Computer Science
  • Journal of Computer Science and Technology
  • 2014
TLDR
An extensive survey of existing bandwidth allocation mechanisms in the literature is presented, covering the scenarios of both public and private clouds, and investigates their underlying design principles, evaluates the trade-off involved in their design choices and summarizes them in a unified design space, with the hope of conveying some meaningful insights for better designs in the future. Expand
Links as a Service (LaaS): Feeling Alone in the Shared Cloud
TLDR
This paper introduces Links as a Service, a new abstraction for cloud service that provides physical isolation of network links that can improve the application performance by up to 200%, at the cost of a 10% reduction in the cloud utilization. Expand
Leveraging tenant flexibility in resource allocation for virtual networks
TLDR
This paper designs a novel resource demand model for tenants to flexibly trade off between application performance and cost, and proposes a work-conserving allocation algorithm, WCA, for deploying virtual networks with time-varying resource demands. Expand
...
1
2
3
4
5
...

References

SHOWING 1-10 OF 53 REFERENCES
Towards predictable datacenter networks
TLDR
The case for extending the tenant-provider interface to explicitly account for the network is made, and the design of virtual network abstractions that capture the trade-off between the performance guarantees offered to tenants, their costs and the provider revenue are proposed. Expand
Flyways To De-Congest Data Center Networks
TLDR
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. Expand
VL2: a scalable and flexible data center network
TLDR
VL2 is a practical network architecture that scales to support huge data centers with uniform high capacity between servers, performance isolation between services, and Ethernet layer-2 semantics, and is built on a working prototype. Expand
Flyways To DeCongest Data Center Networks
  • 2009
A study of application demands from a production datacenter of 1500 servers shows that except for a few outliers, application demands can be generally met by a network that is slightlyExpand
A flexible model for resource management in virtual private networks
TLDR
A new service interface is proposed, termed a hose, to provide the appropriate performance abstraction to manage network resources in the face of increased uncertainty, and the statistical multiplexing and resizing techniques deal effectively with uncertainties about the traffic. Expand
A scalable, commodity data center network architecture
TLDR
This paper shows how to leverage largely commodity Ethernet switches to support the full aggregate bandwidth of clusters consisting of tens of thousands of elements and argues that appropriately architected and interconnected commodity switches may deliver more performance at less cost than available from today's higher-end solutions. Expand
In VINI veritas: realistic and controlled network experimentation
TLDR
VINI's high-level design and the challenges of virtualizing a single network are presented, and an implementation of VINI on PlanetLab, running the "Internet In a Slice", shows that it provides a realistic and controlled environment for evaluating new protocols and services. Expand
A solver for the network testbed mapping problem
TLDR
The design, implementation, and evaluation of a solver for this problem, which is in production use on the Netbed shared network testbed, builds on simulated annealing to find very good solutions in a few seconds for the historical workload, and scales gracefully on large well-connected synthetic topologies. Expand
Rethinking virtual network embedding: substrate support for path splitting and migration
TLDR
This paper simplifies virtual link embedding by allowing the substrate network to split a virtual link over multiple substrate paths and employing path migration to periodically re-optimize the utilization of the substrates network. Expand
SecondNet: a data center network virtualization architecture with bandwidth guarantees
TLDR
This paper proposes virtual data center (VDC) as the unit of resource allocation for multiple tenants in the cloud and introduces a centralized VDC allocation algorithm for bandwidth guaranteed virtual to physical mapping. Expand
...
1
2
3
4
5
...