Volker Hilt

Unreeling netflix: Understanding and improving multi-CDN movie delivery

Netflix is the leading provider of on-demand Internet video streaming in the US and Canada, accounting for 29.7% of the peak downstream traffic in US. Understanding the Netflix architecture and its performance can shed light on how to best optimize its design as well as on the design of similar on-demand streaming services. In this paper, we perform a measurement study of Netflix to uncover its architecture and service strategy. We find that Netflix employs a blend of data centers and Content Delivery Networks (CDNs) for content distribution. We also perform active measurements of the three CDNs employed by Netflix to quantify the video delivery bandwidth available to users across the US. Finally, as improvements to Netflixs current CDN assignment strategy, we propose a measurement-based adaptive CDN selection strategy and a multiple-CDN-based video delivery strategy, and demonstrate their potentials in significantly increasing users average bandwidth.

Greening the internet with content-centric networking

Our energy efficiency analysis of various content dissemination strategies reveals that a change in network architecture from host-oriented to content-centric networking (CCN) can open new possibilities for energy-efficient content dissemination. In this paper, we consider energy-efficient CCN architecture and validate its energy efficiency via trace-based simulations. The results confirm that CCN is more energy efficient than conventional CDNs and P2P networks, even under incremental deployment of CCN-enabled routers.

Controlling overload in networks of SIP servers

The session initiation protocol (SIP) is rapidly being adopted as the signaling protocol for establishing, modifying and terminating multimedia sessions. With the increasing use of SIP in large deployments, it is now becoming apparent that the current SIP design does not easily scale up to large network sizes and SIP servers are not well equipped to handle overload conditions. When a SIP server is operating close to or above its capacity limit, message retransmissions by various SIP timers can cause the network to be severely overloaded and result in an extremely low goodput. In this paper, we first provide a detailed analysis of the behavior of SIP servers under overload. We show that SIP servers are often unable to recover from congestion collapse once it has occurred and that overload can spread throughout a network of SIP servers. We then discuss mechanisms and algorithms for controlling overload in these servers. We found that performing overload control locally at a server provides a simple remedy for light cases of overload; however, it is ineffective in handling higher amounts of load. Finally, we investigate distributed overload control mechanisms for SIP and show that they are effective in controlling overload of SIP servers.

Toward energy-efficient content dissemination

A major role of todays Internet is to provide efficient content dissemination among users, such as distributing multimedia content and sharing user generated data. To meet the ever increasing demands, the Internet has been rapidly growing, and it now includes a web of tens of millions of networked devices ranging from content servers to core and edge routers to home gateways. Due to the sheer numbers, however, it is reported that these devices, such as those used for content delivery, consume a considerable amount of energy. While optimizing the energy efficiency of data centers is well studied in the literature, understanding the energy efficiency of various content dissemination strategies has received comparatively little attention thus far. In this article we review existing content dissemination architectures and survey the energy efficiency of various network devices used for content delivery. The energy efficiency comparison using simple trace-based simulations reveals that a change from a host-oriented to a content-centric networking model can substantially improve energy efficiency of content dissemination. Our preliminary results are encouraging and will stimulate further research in this direction.

On the design of content-centric MANETs

Content-centric networking focuses on data delivery rather than end-to-end reachability by decoupling resources from the hosts they reside on. We consider content-centric networking as a fundamental driver for mobile ad hoc network (MANET) protocol design. We systematically evaluate the suitability and effectiveness of existing approaches toward designing a content-centric MANET. We leverage the extensive prior work on both resource discovery and routing. To examine and compare the various existing designs, we identify a set of representative design alternatives. We develop analytical models for these designs that evaluate their efficiency for a content-centric MANET. Our models provide insights into the strengths and weaknesses of candidate design choices. Our analysis explores the performance boundaries of MANET designs and yields surprising results comparing unstructured flooding to more complex structured solutions. Based on our results, we derive a set of recommendations that are key to the successful design of a content-centric MANET.

Network-aware service placement in a distributed cloud environment

We consider a system of compute and storage resources geographically distributed over a large number of locations connected via a wide-area network. By distributing the resources, latency to users can be decreased, bandwidth costs reduced and availablility increased. The challenge is to distribute services with varying characteristics among the data centers optimally. Some services are very latency sensitive, others need vast amounts of storage, and yet others are computationally complex but do not require hard deadlines on execution. We propose efficient algorithms for the placement of services to get the maximum benefit from a distributed cloud systems. The algorithms need input on the status of the network, compute resources and data resources, which are matched to application requirements. This demonstration shows how a network-aware cloud can combine all three resource types - computation, storage, and network connectivity - in distributed cloud environments. Our dynamic service placement algorithm monitors the network and data center resources in real-time. Our prototype uses the information gathered to place or migrate services to provide the best user experience for a service.

A survey of research on the application-layer traffic optimization problem and the need for layer cooperation

A significant part of Internet traffic today is generated by peer-to-peer applications, used traditionally for file sharing, and more recently for real-time communications and live media streaming. Such applications discover a route to each other through an overlay network with little knowledge of the underlying network topology. As a result, they may choose peers based on information deduced from empirical measurements, which can lead to suboptimal choices. We refer to this as the application layer traffic optimization (ALTO) problem and present a survey of existing literature. We summarize and compare existing approaches, identify open research issues, and state the need for layer cooperation as a solution to the ALTO problem.

Interactions between HTTP adaptive streaming and TCP

HTTP adaptive streaming (HAS) is quickly becoming a popular mechanism for delivering on-demand video content over the Internet. The chunked transmission and application-layer adaptation create a very different traffic pattern than traditional progressive video downloads where the entire video is downloaded with a single request. In this paper, we investigate experimentally the interplay between HAS and the network transport control protocol (TCP). We investigate the impact of network delay on achievable throughput and discover that HAS streams cannot fully utilize the available bandwidth due to the start and stop nature of HAS traffic patterns and its interaction with TCP. We investigate TCP pacing as a potential solution to this issue, particularly for packet losses that occur as a result of bursting packets into the network at the start of a transmission. We find that pacing can significantly increase a TCP flows congestion window but it does not necessarily translate into higher throughput. Instead, we find that packet losses at the end of chunk transmission have a greater impact on throughput.

Waiting for Anonymity: Understanding Delays in the Tor Overlay

Although Tor is the most widely used overlay for providing anonymity services, its users often experience very high delays. Because much of Tor usage is for Web applications, which are sensitive to latency, it is critical to reduce delays in Tor. To take an important step in this direction, we seek an in-depth understanding of delays in Tor. By taking snapshots of the entire Tor network within a short time window, we are able to study the delay distribution of the entire router population. We also monitor delays introduced by individual Tor routers over extended periods of time. Our results indicate that apart from delays introduced by routers, overlay network latency also plays a significant role in delays in Tor. We have also observed that at any time, there exist huge differences in the delays introduced by different routers. Our results reveal key performance characteristics of Tor system behavior and provide valuable insights for improving the Tor performance.

NetServ Framework Design and Implementation 1.0

Eyeball ISPs today are under-utilizing an important asset: edge routers. We present NetServ, a programmable node architecture aimed at turning edge routers into distributed service hosting platforms. This allows ISPs to allocate router resources to content publishers and application service providers motivated to deploy content and services at the network edge. This model provides important benefits over currently available solutions like CDN. Content and services can be brought closer to end users by dynamically installing and removing custom modules as needed throughout the network. Unlike previous programmable router proposals which focused on customizing features of a router, NetServ focuses on deploying content and services. All our design decisions reflect this change in focus. We set three main design goals: a wide-area deployment, a multi-user execution environment, and a clear economic benefit. We built a prototype using Linux, NSIS signaling, and the Java OSGi framework. We also implemented four prototype applications: ActiveCDN provides publisher-specific content distribution and processing; KeepAlive Responder and Media Relay reduce the infrastructure needs of telephony providers; and Overload Control makes it possible to deploy more flexible algorithms to handle excessive traffic.