Zhaogeng Li

Could in-network caching benefit information-centric networking?

Information-Centric Networking is (ICN) [1] gaining increasingly concerns, as an important direction of the Future Internet Architecture research. Although In-network caching is considered as one of the most significant properties of ICN, the cache policy for ICN is still little explored. In this paper, we formulate the in-network caching problem of ICN into Mixed-Integer Linear Programming problem. We also propose a novel cache policy named LB (Least Benefit) with taking into account the distance factor and a new forwarding scheme with shallow flooding (FSF for short) to improve the performance further,. Our simulation results show that with in-networking caching, the average hops of the ICN network can be reduced significantly by nearly 50% with simple cache policy like LFU and with some simple improvement such as LB and FSF the average hop can be reduced further.

A survey on Information-centric Networking: Rationales, designs and debates

The basic function of the Internet is to delivery data (what) to serve the needs of all applications. IP names the attachment points (where) to facilitate ubiquitous interconnectivity as the current way to deliver data. The fundamental mismatch between data delivery and naming attachment points leads to a lot of challenges, e.g., mapping from data name to IP address, handling dynamics of underlying topology, scaling up the data distribution, and securing communication, etc. Information-centric networking (ICN) is proposed to shift the focus of communication paradigm from where to what, by making the named data the first-class citizen in the network, The basic consensus of ICN is to name the data independent from its container (space dimension) and session (time dimension), which breaks the limitation of point-to-point IP semantic. It scales up data distribution by utilizing available resources, and facilitates communication to fit diverse connectivity and heterogeneous networks. However, there are only a few consensuses on the detailed design of ICN, and quite a few different ICN architectures are proposed. This paper reveals the rationales of ICN from the perspective of the Internet evolution, surveys different design choices, and discusses on two debatable topics in ICN, i.e., self-certifying versus hierarchical names, and edge versus pervasive caching. We hope this survey helps clarify some mis-understandings on ICN and achieve more consensuses.

Compression of pending interest table with Bloom filter in content centric network

We propose compressed PIT structure based on Bloom Filter. And we argue that United Bloom Filter is better than Counting Bloom Filter in the PIT compression.

Interest cash: an application-based countermeasure against interest flooding for dynamic content in named data networking

As a design of information-centric network architecture, Named Data Networking (NDN) provides content-based security. The signature binding the name with the content is the key point of content-based security in NDN. However, signing a content will introduce a significant computation overhead, especially for dynamically generated content. Adversaries can take advantages of such computation overhead to deplete the resources of the content provider. In this paper, we propose Interest Cash, an application-based countermeasure against Interest Flooding for dynamic content. Interest Cash requires a content consumer to solve a puzzle before it sends an Interest. The content consumer should provide a solution to this puzzle as cash to get the signing service from the content provider. The experiment shows that an adversary has to use more than 300 times computation resources of the content provider to commit a successful attack when Interest Cash is used.

HTTP-CCN gateway: Adapting HTTP protocol to Content Centric Network

CCN is one of the future Internet architecture. However, the lack of real traffic becomes an obstacle to advanced CCN researches. HTTP, as an content-oriented application-layer protocol working over current Internet, is similar to CCN in many aspects. In this demo, we try to convert HTTP traffic into CCN traffic with HTTP-CCN gateway. Although HTTP is not equivalent to CCN, we can design carefully to make the conversion correct in most situations. The gateway will introduce real traffic into CCN testbed to support CCN researches. Our demonstration will show how to use this gateway.

An Easy Matrix Computation Based Simulator of NDN

This paper presents an easy matrix computation based simulator of NDN network architecture. This simulator is just like Newtonian mechanics in some way, once given the system initial state, routing table, content distribution, initial data requesting information of NDN network, the simulator can figure out all of the subsequent network state based on matrix operation. This simulator turns packet process event, such as interest generating, interest forwarding, content hitting and transmitting, to matrix computation on the network scale. This simulator is much easier to use than NS3 and CCNx, but it can provide similar simulation result and more details. And the computational complexity depends on number of involved contents and nodes, which makes analysis between small group of nodes and contents very simple and fast.

EAalo: Enhanced coflow scheduling without prior knowledge in a datacenter network

Coflow scheduling without prior knowledge has been proposed recently. However, the previous solution, Aalo, has two problems: it does not explicitly control the flow rate; it assumes the network is ideally non-blocking (with perfect traffic balancing and no oversubscription). In this paper, we show the performance loss caused by the above two problems. We propose EAalo to cope with the problems. It has three enhancements to Aalo: per-flow bandwidth enforcement, centralized traffic balancing and oversubscription adaption. We evaluate EAalo with simulations. The simulation results show that EAalo has much better performance than Aalo. It can speed up coflow completion by up to 11%, 17% and 19% in non-blocking networks, networks without oversubscription and networks with 2∶1 oversubscription respectively.

P4DB: On-the-fly debugging of the programmable data plane

While extending network programmability to a larger degree, P4 also raises the risks of incurring runtime bugs after the deployment of P4 programs. These runtime bugs, if not handled promptly and properly, can ruin the functionality and performance of networks. Unfortunately, the absence of runtime debuggers makes troubleshooting of P4 program bugs challenging and intricate for operators. This paper is devoted to the on-the-fly debugging of runtime bugs in P4-enabled networks. We propose P4DB, a general debugging platform that empowers operators to debug P4 programs in three levels of visibility by provisioning operator-friendly primitives. By P4DB, operators can use the watch primitive to quickly narrow the debugging scope from network level or device level to table level, then use the break and next primitives to decompose the match-action table into three steps and troubleshoot the runtime bugs step by step. We implemented a prototype of P4DB and evaluated the performance in terms of the data plane, control plane and control channel. On P4-specific programmable data plane, P4DB merely introduces a small throughput penalty (1.3%∼13.8%) and imposes a little-increased delay (0.6%∼11.9%).

VMS: Traffic balancing based on virtual switches in datacenter networks

There have been many traffic balancing solutions for datacenter networks. All of them require modifications to the network fabric or/and virtual machines. In this paper, we propose Virtual Multi-channel Scatter (VMS), a new traffic balancing solution in datacenter networks. VMS works in the virtual switches between the network fabric and virtual machines. It can be deployed by datacenter operators at a relatively low cost without extra restrictions to virtual machine users. VMS scatters packets in one TCP flow to several different forwarding paths. It employs an adaptive path selection based on the virtual window size of different paths. We implemented VMS based on OVS. Our evaluation demonstrates that VMS improves traffic balancing very well, and the performance of VMS is approximate to MPTCP in almost all the cases, while only modifies virtual switches. Further, the overhead of VMS is tolerable.