Data center networking with multipath TCP
read more
Citations
DevoFlow: scaling flow management for high-performance networks
Design, implementation and evaluation of congestion control for multipath TCP
DeTail: reducing the flow completion time tail in datacenter networks
On the impact of packet spraying in data center networks
Decentralized task-aware scheduling for data center networks
References
A scalable, commodity data center network architecture
VL2: a scalable and flexible data center network
A study of non-blocking switching networks
BCube: a high performance, server-centric network architecture for modular data centers
Hedera: dynamic flow scheduling for data center networks
Related Papers (5)
Frequently Asked Questions (16)
Q2. What are the contributions in "Data center networking with multipath tcp" ?
The authors propose a natural evolution of data center transport from TCP to multipath TCP. The authors show that multipath TCP can effectively and seamlessly use available bandwidth, providing improved throughput and better fairness in these new topologies when compared to single path TCP and randomized flow-level load balancing. The authors also show that multipath TCP outperforms laggy centralized flow scheduling without needing centralized control or additional infrastructure.
Q3. What is the way to schedule a flow?
during a scheduling period, a flow’s average throughput is greater than 10% of the interface speed, it is explicitly scheduled.
Q4. What is the main advantage of using a hierarchical topology?
Traditionally data centers have been built using hierarchical topologies: racks of hosts connect to a top-of-rack switch; these switches connect to aggregation switches; in turn these are connected to a core switch.
Q5. What is the default method for randomised load balancing?
Solutions such as ECMP or multiple VLANs provide the basis for randomised load balancing as the default path selection mechanism.
Q6. What is the simplest solution to the traffic concentration problem?
The simplest solution is to use randomized load balancing, where each flow is assigned a random path from the set of possible paths.
Q7. What is the advantage of this approach?
The great advantage of this approach is that the linked congestion controller in each MPTCP end system can act on very short timescales to move its own traffic from paths it observes to be more congested, onto paths it observes to be less congested.
Q8. What is the reason for the growth in data centers?
Recent growth in cloud applications from companies such as Google, Microsoft, and Amazon has resulted in the construction of data centers of unprecedented size.
Q9. What is the definition of a flow scheduling algorithm?
Host-limited Flows Hedera’s flow scheduling algorithm is based on the assumption that long-lived flows contribute most of the bytes and therefore it only needs to schedule those flows.
Q10. What are the main advantages of MPTCP?
First the authors wish to understand the potential benefits of MPTCP with respect to the three major topologies in the literature: FatTree, VL2 and BCube.
Q11. What is the intuition for scheduling big flows?
The intuition is that if the authors only schedule the big flows the authors can fully utilize all the bandwidth, and yet have a small scheduling cost, as dictated by the small number of flows.
Q12. How many subflows are needed for a fattree network?
This might seem like quite a high number, but for an 8192-node FatTree network there are 256 distinct paths between each host pair, so only a small fraction of the paths are needed to achieve full utilization.
Q13. How many subflows are needed to fully utilize the network?
Although such bottleneck links are loadbalanced, with FatTree in particular, other links cannot be fully utilized, and it takes more than two subflows to spread load across sufficient paths to fully utilize the network.
Q14. What are the differences between FatTree and VL2?
They differ in that FatTree uses larger quantities of lower speed (1Gb/s) links between switches, whereas VL2 uses fewer faster (10Gb/s) links.
Q15. What is the problem with dense interconnects?
while such dense interconnects can in principle support the full cross-sectional bandwidth of every host communicating flat out simultaneously, the denseness of interconnection poses a difficult challenge for routing.
Q16. What is the way to analyze traffic in the three topologies?
In all three topologies, with the right path selection this traffic pattern should just be able to load the network to full capacity but no more.