Rohan Gandhi

Dynamic scheduling of network updates

We present Dionysus, a system for fast, consistent network updates in software-defined networks. Dionysus encodes as a graph the consistency-related dependencies among updates at individual switches, and it then dynamically schedules these updates based on runtime differences in the update speeds of different switches. This dynamic scheduling is the key to its speed; prior update methods are slow because they pre-determine a schedule, which does not adapt to runtime conditions. Testbed experiments and data-driven simulations show that Dionysus improves the median update speed by 53--88% in both wide area and data center networks compared to prior methods.

Duet: cloud scale load balancing with hardware and software

Load balancing is a foundational function of datacenter infrastructures and is critical to the performance of online services hosted in datacenters. As the demand for cloud services grows, expensive and hard-to-scale dedicated hardware load balancers are being replaced with software load balancers that scale using a distributed data plane that runs on commodity servers. Software load balancers offer low cost, high availability and high flexibility, but suffer high latency and low capacity per load balancer, making them less than ideal for applications that demand either high throughput, or low latency or both. In this paper, we present Duet, which offers all the benefits of software load balancer, along with low latency and high availability -- at next to no cost. We do this by exploiting a hitherto overlooked resource in the data center networks -- the switches themselves. We show how to embed the load balancing functionality into existing hardware switches, thereby achieving organic scalability at no extra cost. For flexibility and high availability, Duet seamlessly integrates the switch-based load balancer with a small deployment of software load balancer. We enumerate and solve several architectural and algorithmic challenges involved in building such a hybrid load balancer. We evaluate Duet using a prototype implementation, as well as extensive simulations driven by traces from our production data centers. Our evaluation shows that Duet provides 10x more capacity than a software load balancer, at a fraction of a cost, while reducing latency by a factor of 10 or more, and is able to quickly adapt to network dynamics including failures.

Fast rendezvous for multiple clients for cognitive radios using coordinated channel hopping

A primary challenge in exploiting Cognitive Radio Networks (CRNs), known as the rendezvous problem, is for the users to find each other in the dynamic open spectrum. We study blind rendezvous, where users search for each other without any infrastructural aid. Previous work in this area have focused on efficient blind rendezvous algorithms for two users but the solution for multiple users is still far from optimal. In particular, when two users encounter, one user inherits the others hopping sequence but the sequence is never shortened or split among the encountering users. We denote this class of algorithms as uncoordinated channel hopping algorithms. In this paper, we introduce a new class of distributed algorithms for multi-user blind rendezvous, called Coordinated Channel Hopping (CCH), where users adjust, or coordinate, the sequence of channels being hopped as they rendezvous pairwise. Compared to existing rendezvous algorithms, our algorithms achieve 80% lower Time To Rendezvous (TTR) in case of multiple users.

Mercury: bringing efficiency to key-value stores

While the initial wave of in-memory key-value stores has been optimized for serving relatively fixed content to a very large number of users, an emerging class of enterprise-scale data analytics workloads focuses on capturing, analyzing, and reacting to data in real-time. At the same time, advances in network technologies are shifting the performance bottleneck from the network to the memory subsystem. To address these new trends, we present a bottom-up approach to building a high performance in-memory key-value store, Mercury, for both traditional, read-intensive as well as emerging workloads with high write-to-read ratio. Mercurys architecture is based on two key design principles: (i) economizing the number of DRAM accesses per operation, and (ii) reducing synchronization overheads. We implement these principles with a simple hash table with linked-list based chaining, and provide high concurrency with a fine-grained, cache-friendly locking scheme. On a commodity single-socket server with 12 cores, Mercury scales with number of cores and executes 14 times more queries/second than a popular hash-based key-value system, Memcached, for both read and write-heavy workloads.

Yoda: a highly available layer-7 load balancer

Layer-7 load balancing is a foundational building block of online services. The lack of offerings from major public cloud providers have left online services to build their own load balancers (LB), or use third-party LB design such as HAProxy. The key problem with such proxy-based design is each proxy instance is a single point of failure, as upon its failure, the TCP flow state for the connections with the client and server is lost which breaks the user flows. This significantly affects user experience and online services revenue. In this paper, we present Yoda, a highly available, scalable and low-latency L7-LB-as-a-service in a public cloud. Yoda is based on two design principles we propose for achieving high availability of a L7 LB: decoupling the flow state from the LB instances and storing it in a persistent storage, and leveraging the L4 LB service to enable each L7 LB instance to use the virtual IP in interacting with both the client and the server (called front-and-back indirection). Our evaluation of Yoda prototype on a 60-VM testbed in Windows Azure shows the overhead of decoupling TCP state into a persistent storage is very low (<1 msec), and Yoda maintains all flows during LB instance failures, addition, removal, as well as user policy updates. Our simulation driven by a one-day trace from production online services show that compared to using Yoda by each tenant, Yoda-as-a-service reduces L7 LB instance cost for the tenants by 3.7x while providing 4x more redundancy.

The impact of inter-layer network coding on the relative performance of MRC/MDC WiFi media delivery

A primary challenge in multicasting video in a wireless LAN is to deal with the client diversity -- clients may have different channel characteristics and hence receive different numbers of transmissions from the AP. A promising approach to overcome this problem is to combine scalable video coding techniques such as MRC or MDC, which divide a video stream into multiple substreams, with inter-layer network coding. The fundamental challenge in such an approach is to determine the strategy of coding the packets across different layers that maximizes the number of decoded layers at all clients. In [7], the authors showed that inter-layer NC indeed helps the delivery of MRC coded media over the WiFi, and proposed how to efficiently search for the optimal coding strategies online. In this paper, we study (1) how NC can help with WiFi delivery of MDC media, and (2) in particular, due to the different decoding requirements of MDC from MRC, whether WiFi delivery of MDC media can benefit more from NC compared to that of MRC media. Our simulation results are somewhat surprising. Even though MDC is generally shown to outperform MRC in lossy channels, most of the benefit of MDC over MRC is lost after applying NC to both schemes.

Catalyst: Unlocking the Power of Choice to Speed up Network Updates

Speeding up network updates is crucial to maintain high agility and to react quickly to network failures. In this paper, we present Ctalyst---a new design to reduce the network update time. We observe that networks offer a power of choice, where there are many equally-good alternative paths that traffic flows can be assigned to, which is facilitated by redundancy in networks. Catalyst exploits this power of choice to assign flows to alternative paths to merge stages in the dependency graph (that captures the update plan), which in turn reduces the total update time. Furthermore, we observe that because of the prevalence of switch stragglers---switches that unexpectedly take longer time to update, simply assigning a flow to a single (shortest) path is not an optimal design as even a single switch straggler can substantially increase the update time. Thus, the second principle in Catalyst is to compute multiple paths for individual flows offline, among which one would be selected at runtime based on temporal switch conditions, in order to enable a fast update. Our evaluation using a load-balancer setting in a data center network shows that Catalyst effectively reduces the total update time by 1.14--2.15x.

Saath: Speeding up CoFlows by Exploiting the Spatial Dimension

CoFlow scheduling improves data-intensive application performance by improving their networking performance. State-of-the-art CoFlow schedulers in essence approximate the classic online Shortest-Job-First (SJF) scheduling, designed for a single CPU, in a distributed setting, with no coordination among how the flows of a CoFlow at individual ports are scheduled, and as a result suffer two performance drawbacks: (1) The flows of a CoFlow may suffer the out-of-sync problem -- they may be scheduled at different times and become drifting apart, negatively affecting the CoFlow completion time (CCT); (2) FIFO scheduling of flows at each port bears no notion of SJF, leading to suboptimal CCT. We propose Saath, an online CoFlow scheduler that overcomes the above drawbacks by explicitly exploiting the spatial dimension of CoFlows. In Saath, the global scheduler schedules the flows of a CoFlow using an all-or-none policy which mitigates the out-of-sync problem. To order the CoFlows within each queue, Saath resorts to a Least-Contention-First (LCoF) policy which we show extends the gist of SJF to the spatial dimension, complemented with starvation freedom. Our evaluation using an Azure testbed and simulations of two production cluster traces show that compared to Aalo, Saath reduces the CCT in median (P90) cases by 1.53x (4.5x) and 1.42x (37x), respectively.

Multicasting MDC videos to receivers with different screen resolution

A primary challenge in multicasting video in a wireless LAN has been to deal with the client diversity in terms of channel diversity: clients may have different channel characteristics and hence receive different numbers of packet transmissions from the AP. Various schemes exploiting layered video coding schemes such as MRC and MDC have been proposed to address this problem. With the advent of smartphones, a new form of client heterogeneity, that different portable devices have different screen resolution and hence desire different numbers of layers, has become increasingly prominent. In this paper, we propose a practical transmission strategy selection method that takes into consideration this new form of client diversity and show it can increase the number of layers received by clients by up to 62%, compared to a scheme that is oblivious of the client screen resolution diversity.