Massimo Gallo

Optimal multipath congestion control and request forwarding in Information-Centric Networks

The evolution of the Internet into a distributed Information access system calls for a paradigm shift to enable an evolvable future network architecture. Information-Centric Networking (ICN) proposals rethink the communication model around named data, in contrast with the host-centric transport view of TCP/IP. Information retrieval is natively pull-based, driven by user requests, point-to-multipoint and intrinsically coupled with in-network caching. In this paper, we tackle the problem of joint multipath congestion control and request forwarding in ICN for the first time. We formulate it as a global optimization problem with the twofold objective of maximizing user throughput and minimizing overall network cost. We solve it via decomposition and derive a family of optimal congestion control strategies at the receiver and of distributed algorithms for dynamic request forwarding at network nodes. An experimental evaluation of our proposal is carried out in different network scenarios to assess the performance of our design and to highlight the benefits of an ICN approach.

P2P-TV Systems under Adverse Network Conditions: A Measurement Study

In this paper we define a simple experimental set- up to analyze the behavior of commercial P2P-TV applications under adverse network conditions. Our goal is to reveal the ability of different P2P-TV applications to adapt to dynamically changing conditions, such as delay, loss and available capacity, e.g., checking whether such systems implement some form of congestion control. We apply our methodology to four popular commercial P2P-TV applications: PPLive, SOPCast, TVants and TVUPlayer. Our results show that all the considered applications are in general capable to cope with packet losses and to react to congestion arising in the network core. Indeed, all applications keep trying to download data by avoiding bad paths and carefully selecting good peers. However, when the bottleneck affects all peers, e.g., it is at the access link, their behavior results rather aggressive, and potentially harmful for both other applications and the network.

On the performance of bandwidth and storage sharing in information-centric networks

Information Centric Network (ICN) proposals, have recently emerged to define new network architectures where content, and not its location, becomes the core of the communication model. Such new paradigms push data storage and delivery at network layer and are designed to cope with current Internet usage, mainly centered around content dissemination and retrieval.In this paper, we develop an analytical model of ICN storage and bandwidth sharing under fairly general assumptions on total demand, topology, content popularity and limited network resources. Our study applies to a class of content oriented networks identified by receiver-driven packet-based communication with in-network storage. We derive a closed-form expression for expected stationary delivery time as a function of hit/miss probabilities at network caches, content popularity and cache sizes. Our analytical results, supported by packet level simulations, can be used to analyze fundamental trade-offs of ICN architectures. They also provide an essential building block for the design and evaluation of ICN protocols.

Performance evaluation of the random replacement policy for networks of caches

The overall performance of content distribution networks as well as recently proposed information-centric networks rely on both memory and bandwidth capacities. The hit ratio is the key performance indicator which captures the bandwidth/memory tradeoff for a given global performance. This paper focuses on the estimation of the hit ratio in a network of caches that employ the Random replacement policy (RND). Assuming that requests are independent and identically distributed, general expressions of miss probabilities for a single RND cache are provided as well as exact results for specific popularity distributions (such results also hold for the FIFO replacement policy). Moreover, for any Zipf popularity distribution with exponent @a>1, we obtain asymptotic equivalents for the miss probability in the case of large cache size. We extend the analysis to networks of RND caches, when the topology is either a line or a homogeneous tree. In that case, approximations for miss probabilities across the network are derived by neglecting time correlations between miss events at any node; the obtained results are compared to the same network using the Least-Recently-Used discipline, already addressed in the literature. We further analyze the case of a mixed tandem cache network where the two nodes employ either Random or Least-Recently-Used policies. In all scenarios, asymptotic formulas and approximations are extensively compared to simulation results and shown to be very accurate. Finally, our results enable us to propose recommendations for cache replacement disciplines in a network dedicated to content distribution.

Hierarchical Content Stores in High-Speed ICN Routers: Emulation and Prototype Implementation

Recent work motivates the design of Information-centric routers that make use of hierarchies of memory to jointly scale in the size and speed of content stores. The present paper advances this understanding by (i) instantiating a general purpose two-layer packet-level caching system, (ii) investigating the solution design space via emulation, and (iii) introducing a proof-of-concept prototype. The emulation-based study reveals insights about the broad design space, the expected impact of workload, and gains due to multi-threaded execution. The full-blown system prototype experimentally confirms that, by exploiting both DRAM and SSD memory technologies, ICN routers can sustain cache operations in excess of 10Gbps running on off-the-shelf hardware.

Pending Interest Table Sizing in Named Data Networking

Named Data Networking (NDN) has emerged as a promising candidate for shifting Internet communication model from host-centric to content-centric. A core component of NDN is its stateful forwarding plane: Content Routers keep track of pending requests (Interests) storing them in dedicated tables at routers (Pending Interest Tables). A thorough analysis of PIT scalability is fundamental for deploying NDN as a whole and questions naturally arise about memory requirements and feasibility at wire-speed. While previous works focus on data structures design under the threat of PIT state explosion, we develop for the first time an analytical model of PIT dynamics as a function of relevant system parameters. We provide a closed form characterization of average and maximum PIT size value at steady state. We build an experimental platform with high speed content router implementation to investigate PIT dynamics and to confirm the accuracy of our analytical findings. Finally, we provide guidelines on optimal PIT dimensioning and analyze the case of an ISP aggregation network with a trace-driven packet delay distribution. We conclude that, even in absence of caching and under optimal network bandwidth usage, PIT size results to be small in typical network settings.

Scalable mobile backhauling via information-centric networking

The rapid traffic growth fueled by mobile devices spread and high speed network access calls for substantial innovation at network layer. The content-centric nature of Internet usage highlights the limitations of the host-centric model in coping with dynamic content-to-location binding, mobility, multicast, multi-homing, etc. If transmission capacity speedups in the backhaul may hide inefficiencies in the short term, the hostcentric communication model needs to be revisited to sustain future mobile demand. In this paper, we first identify and quantify the opportunities for backhaul evolution by analyzing a large set of traffic measurements collected between mobile core and backhaul of Orange France. The analysis reveals that 50% of HTTP requests are cacheable and traffic can be reduced from 60% to 95% during the peak hour by using 350GBs to 1TB of memory overall. Motivated by such significant opportunities for latency reduction and network cost savings, we present a solution based on Information-Centric Networking (ICN). First results of a large scale experimentation with 100 Linux servers and customized software, in a realistic network setting, provide a glimpse into ICN gains even under naive caching: a factor three reduction in delivery time and almost 40% bandwidth savings, when compared to existing alternatives.

CliMB: Enabling Network Function Composition with Click Middleboxes

Click has significant advantages for middlebox development, including modularity, extensibility, and reprogrammability. Despite these features, Click still has no native TCP support and only uses nonblocking I/O, preventing its applicability to middleboxes that require access to L7 data and blocking I/O. We bridge this gap by introducing Click middleboxes (CliMB). CliMB provides a full-fledged modular TCP layer supporting congestion control, TCP options, both blocking and nonblocking I/O, as well as socket and zero-copy APIs to applications. As a result, any TCP network function may now be realized in Click. As proof of concept, we develop a zero-copy SOCKS proxy server using CliMB that shows up to 4x gains compared to an equivalent implementation using the Linux in-kernel network stack.

A programmable data plane for heterogeneous NFV platforms.

Network function virtualization (NFV) has recently allowed the rapid deployment of network functions. As software implementations become the main option for NFV, performance requirements call for an increased level of hardware support and acceleration. Additionally, the higher demand for flexibility and optimization also requires protocol stack customization at different layers. To fulfill these requirements, we introduce in this paper a programmable data plane (PDP) whose goals are to enable on-the-fly customization of L2-L7 stacks and to integrate both general-purpose CPUs and hardware accelerators. The PDP data plane provides a set of modular elements that the control plane instantiates and orchestrates to compose L2-L7 network functions, allowing a centralized controller to reconfigure heterogeneous network devices and manage state across network functions. We present examples of how PDP data plane modules can be used to realize different functions, such as an information-centric networking (ICN) router and an HTTP reverse proxy. Our preliminary results show that hardware acceleration provides significant benefits at negligible cost, and that careful use of modularity brings a minimal latency penalty, while providing high flexibility and reconfigurability to the network.

Optimal multipath congestion control and request forwarding in information-centric networks

In this paper we consider the problem of joint congestion control and request forwarding in Information-Centric Networks, namely the named-data networking architecture (NDN). The network architecture we consider is based on information retrieval natively pull-based, driven by user requests, point-to-multipoint and intrinsically coupled with in-network caching. We formalize the problem as global optimization with non-linear objectives and linear constraints with the twofold objective of maximizing user throughput and minimizing overall network cost. We solve it via decomposition and derive a family of optimal congestion control strategies at the receiver and of distributed algorithms for dynamic request forwarding at network nodes. An experimental evaluation of our proposal is carried out in different network scenarios using realistic workloads, to assess the performance of our design and to highlight the benefits of an ICN approach. The experimentation is carried out using the NDN software router implementation on a large grid infrastructure deployed to enable experimental research.