Check out the new USENIX Web site. next_inactive up previous


iPlane: An Information Plane for Distributed Services

Harsha V. Madhyastha
Tomas Isdal
Michael Piatek
Colin Dixon
Thomas Anderson
Arvind Krishnamurthy
Department of Computer Science and Engineering, University of Washington

Arun Venkataramani
Department of Computer Science, University of Massachusetts Amherst

Abstract

In this paper, we present the design, implementation, and evaluation of iPlane, a scalable service providing accurate predictions of Internet path performance for emerging overlay services. Unlike the more common black box latency prediction techniques in use today, iPlane adopts a structural approach and predicts end-to-end path performance by composing the performance of measured segments of Internet paths. For the paths we observed, this method allows us to accurately and efficiently predict latency, bandwidth, capacity and loss rates between arbitrary Internet hosts. We demonstrate the feasibility and utility of the iPlane service by applying it to several representative overlay services in use today: content distribution, swarming peer-to-peer filesharing, and voice-over-IP. In each case, using iPlane's predictions leads to improved overlay performance.

1 Introduction

The Internet by design is opaque to its applications, providing best effort packet delivery with little or no information about the likely performance or reliability characteristics of different paths. While this is a reasonable design for simple client-server applications, many emerging large-scale distributed services depend on richer information about the state of the network. For example, content distribution networks like Akamai [1], Coral [16], and CoDeeN [52] re-direct each client to the replica providing the best performance for that client. Likewise, voice-over-IP systems such as Skype [45] use relay nodes to bridge hosts behind NAT/firewalls, the selection of which can dramatically affect call quality [39]. Peer-to-peer file distribution, overlay multicast, distributed hash tables, and many other overlay services can benefit from peer selection based on different metrics of network performance such as latency, available bandwidth, and loss rate. Finally, the Internet itself can benefit from more information about itself, e.g., ISPs can monitor the global state of the Internet for reachability and root cause analysis, routing instability, and onset of DDoS attacks.

If Internet performance were easily predictable, its opaqueness might be an acceptable state of affairs. However, Internet behavior is well-known to be fickle, with local hot spots, transient (and partial) disconnectivity, and triangle inequality violations all being quite common [41,2]. Many large-scale services adapt to this state of affairs by building their own proprietary and application-specific information plane. Not only is this redundant, but it prevents new applications from leveraging information already gathered by other applications. The result is often sub-optimal. For example, most implementations of the file distribution tool BitTorrent choose peers at random (or at best using round trip latency estimates); since downloads are bandwidth-dependent, this can yield suboptimal download times. By some estimates, BitTorrent accounts for roughly a third of backbone traffic [37], so inefficiency at this scale is a serious concern. Moreover, implementing an information plane is often quite subtle, e.g., large-scale probing of end-hosts can raise intrusion alarms in edge networks as the traffic can resemble a DDoS attack. This is the most common source of complaints on PlanetLab [38].

To address this, several research efforts, such as IDMaps [15], GNP [34], Vivaldi [11], Meridian [54], and PlanetSeer [55] have investigated providing a common measurement infrastructure for distributed applications. These systems provide only a limited subset of the metrics of interest, most commonly latency between a pair of nodes, whereas most applications desire richer information such as loss rate and bandwidth. Second, by treating the Internet as a black box, most of these services abstract away network characteristics and atypical behavior--exactly the information of value for troubleshooting as well as improving performance. For example, the most common latency prediction methods use metric embeddings which are fundamentally incapable of predicting detour paths as such paths violate the triangle inequality [41,56]. More importantly, being agnostic to network structure, they cannot pinpoint failures, identify causes of poor performance, predict the effect of network topology changes, or assist applications with new functionality such as multipath routing.

In this paper, we move beyond mere latency prediction and develop a service to automatically infer sophisticated network behavior. We develop an Information Plane (iPlane) that continuously performs measurements to generate and maintain an annotated map of the Internet with a rich set of link and router attributes. iPlane uses structural information such as the router-level topology and autonomous system (AS) topology to predict paths between arbitrary nodes in the Internet. The path predictions are combined with measured characteristics of path segments to predict end-to-end path properties for a number of metrics such as latency, available bandwidth, and loss rate. iPlane can also analyze isolated anomalies or obtain a global view of network behavior by correlating observations from different parts of the Internet.

iPlane is designed as a service that distributed applications can query to obtain information about network conditions. Deploying iPlane as a shared service (as opposed to providing a library) has several benefits. First, a common iPlane can exploit temporal and spatial locality of queries across applications to minimize redundant measurement overhead. Second, iPlane can selectively refresh its knowledge of the IP address space based on real query workloads. More generally, iPlane can assimilate measurements made on behalf of all of its clients as well as incorporate information reported by clients to develop a more comprehensive model of Internet behavior over time. We note that all of these arguments have been recognized before [48,8,53,23], however a convincing validation has remained starkly absent.

Our primary contribution is in demonstrating the feasibility of a useful iPlane, e.g., we can infer with high accuracy an annotated map of the Internet every six hours with approximately $100$Kbps of measurement traffic per PlanetLab node. In addition, we develop:

$\bullet$
A common structural model to predict path properties.

$\bullet$
A measurement infrastructure that is deployed on every active PlanetLab site and almost a thousand traceroute and Looking Glass server vantage points (with a lower intensity of probing).

$\bullet$
A toolkit for using BitTorrent swarms to measure links.

$\bullet$
Case studies of popular systems such as CDNs, peer-to-peer file swarming, and VoIP. We show measurable benefits of using iPlane for each of these applications.

iPlane is a modest step towards the vision of a knowledge plane pioneered by Clark et al. [8]. iPlane supplies information about the network and leaves the task of adapting or repairing to the client. Nevertheless, the collection, analysis, and distribution of Internet-scale measurement information is itself a challenging systems engineering problem and the focus of this paper. The goal of gathering a complete picture of the Internet has been recognized earlier in [48]. Our goal is more modest--to gather a coarse-grained map of the Internet sufficient to be of utility in improving overlay performance.


2 Design

We start by discussing the requirements of an Information Plane for distributed services before presenting our design that meets these requirements.

$\bullet$
Accuracy: iPlane should accurately estimate a rich set of performance metrics such as latency, loss-rate, capacity, and available bandwidth.
$\bullet$
Wide coverage: iPlane must predict the performance of arbitrary Internet paths. Many currently deployed prediction services, such as RON [2] and $S^3$ [29], limit their focus to intra-overlay paths.
$\bullet$
Scalability: iPlane should not impose an undue communication load on its measurement infrastructure.
$\bullet$
Unobtrusiveness: Active probes of end-hosts must be coordinated and performed in an unobtrusive manner in order to minimize the possibility of raising intrusion detection alarms.


Table 1: A summary of techniques used in iPlane.
Technique Description Goal Section

generate probe
targets

Obtain prefixes from Routeview's BGP snapshot and cluster groups of prefixes with similar routes.

coverage,
scalability
Section 2.2.1

traceroutes from
vantage points

PlanetLab nodes probe all targets, while Traceroute/Looking Glass servers issue probes to a small subset of the targets.

map topology,
capture path diversity
Section 2.2.1

cluster network
interfaces

Identify network interfaces that are in the same AS and geographically colocated.

build structured topology,
scalability
Section 2.2.2

frontier algorithm

Schedule measurements of link attributes to PlanetLab nodes such that each link is probed by the vantage point closest to it.

accuracy,
balance load
Section 2.3.1

measure link
attributes

PlanetLab nodes measure the loss rate, capacity, and available bandwidth over a subset of paths in the Internet core.

annotate topology
Section 2.3.2

opportunistic
measurements

Leverage existing applications to discover the structure and performance of edge networks.

minimize obtrusiveness,
access link properties
Section 2.4

route composition

Compose segments of observed or reported paths to predict end-to-end paths between a pair of nodes.

path prediction,
performance prediction
Section 2.5


2.1 Overview

iPlane is designed to be deployed as an application-level overlay network with the overlay nodes collectively coordinating the task of generating and maintaining an ``atlas'' of the Internet. The atlas is both extensive and detailed--it comprises the topology of the Internet core and the core's connectivity to representative targets in the edge networks, complete with a rich set of static attributes (such as link delay and link capacity), and recent observations of dynamic properties (such as routes between network elements, path loss rates, and path congestion). iPlane uses systematic active measurements to determine the attributes of the core routers and the links connecting them. In addition, the system performs opportunistic measurements by monitoring actual data transfers to/from end-hosts participating in BitTorrent swarms, thereby exposing characteristics of the edge of the network that typically cannot be obtained from one-way probing, e.g., capacities of access links.

Since it is impractical to probe every Internet end-host to generate the atlas, we cluster end-hosts on the basis of BGP atoms [4]. We approximate a client's performance by a representative target in the same atom as the client. If the client desires greater prediction accuracy, it can voluntarily perform some probes and contribute the paths that it discovers to iPlane; multi-homed clients can benefit from such an operational model. iPlane uses its collected repository of observed paths to predict end-to-end paths between any pair of end-hosts. This prediction is made by carefully composing partial segments of known Internet paths so as to exploit the similarity of Internet routes [31], i.e., routes from two nearby sources tend to converge when heading to the same destination. iPlane predicts a path by splicing a short path segment from the source to an intersection point from which a path going to the destination has been observed in the atlas. To determine intersections between paths, we cluster interfaces that are owned by the same AS and reside in the same PoP, and deem two paths to have intersected if they pass through the same cluster.

Once a path is predicted, iPlane simply composes the measured properties of the constituent path segments to predict the performance of the composite path. For instance, to make a latency prediction, iPlane simply adds the latencies associated with the individual path segments. Or, to predict the end-to-end bandwidth, iPlane computes the minimum of the bandwidth measured of each of the inter-cluster links along the predicted path, and the bandwidth of the client's access link, if available.

The rest of this section describes the techniques used to develop a functional iPlane that has wide coverage, incurs modest measurement load without unduly sacrificing coverage or detail, and uses topology structuring techniques to enable efficient measurement and accurate inference. The techniques are summarized in Table 1.


2.2 Mapping the Internet Topology

iPlane requires geographically distributed vantage points to map the Internet topology and obtain a collection of observed paths. PlanetLab servers, located at over $300$ sites around the world, serve as the primary vantage points. We also enlist the use of public Looking Glass/Traceroute servers for low-intensity probing. Further, we are currently exploring the option of using data from DIMES [43], a system for aggregating low intensity measurements from normal PCs. Our primary tool for determining the Internet topology is traceroute, which allows us to identify the network interfaces on the forward path from the probing entity to the destination. (On PlanetLab, we use an optimized version of the tool to reduce measurement load.) Determining what destinations to probe and how to convert the raw output of traceroute to a structured topology is nontrivial, an issue we address next.


2.2.1 Probe Target Selection

BGP snapshots, such as those collected by RouteViews [33], are a good source of probe targets. iPlane achieves wide coverage for the topology mapping process by obtaining the list of all globally routable prefixes in BGP snapshots, and choosing within each prefix a target .1 address that responds to either ICMP or UDP probes. A .1 address is typically a router and is hence more likely to respond to probes than arbitrary end-hosts.

To reduce measurement load, iPlane clusters IP prefixes into BGP atoms [4] for generating the target list. A BGP atom is a set of prefixes, each of which has the same AS path to it from any given vantage point. BGP atoms can be regarded as representing the knee of the curve with respect to measurement efficiency--probing within an atom might find new routes, but it is less likely to do so [4]. This task of determining a representative set of IP addresses is performed relatively infrequently, typically once every two weeks.

iPlane uses the PlanetLab nodes to perform exhaustive and periodic probing of the representative targets. In addition, iPlane schedules probes from public traceroute servers to a small random set of BGP atoms, typically making a few tens of measurements during the course of a day. The public traceroute servers serve as a valuable source of information regarding local routing policies. Note that in the long run, a functioning iPlane may actually serve to decrease the load on the public traceroute servers as iPlane, rather than the traceroute servers themselves, can be consulted for information on the Internet topology.



2.2.2 Clustering of Interfaces

Traceroute produces a list of network interfaces on the path from source to destination. However, interfaces on the same router, or in the same PoP, may have similar behavior. Hence, we partition network interfaces into ``clusters'' and use this more compact topology for more in-depth measurements and predictions. We define the clusters to include interfaces that are similar from a routing and performance perspective, i.e., interfaces belonging to the same PoP and interfaces within geographically nearby portions of the same AS [46]. Note that this clustering is performed on network interfaces in the Internet core, whereas the clustering of prefixes into BGP atoms was performed for end-host IP addresses. In fact, clustering addresses in the same prefix will be ineffective in the core as geographically distant interfaces are often assigned addresses in the same prefix.

First, iPlane identifies interfaces that belong to the same router. Interfaces that are potential alias candidates are identified using two different techniques. Employing the Mercator [19] technique, UDP probes are sent to a high-numbered port on every router interface observed in traceroutes. Interfaces that return responses with the same source address are considered as possible aliases. In addition, candidate alias pairs are also identified using the fact that interfaces on either end of a long-distance link are usually in the same /30 prefix. Candidate pairs that respond with similar IP-ID values to the UDP probes, and also respond with similar TTLs to the ICMP probes are deemed to be aliases. In one of our typical runs, of the 396,322 alias candidate pairs yielded by the Mercator technique, 340,580 pairs were determined to be aliases. The 918,619 additional alias candidate pairs obtained using the /30 heuristic yielded another 320,150 alias pairs.

Second, iPlane determines the DNS names assigned to as many network interfaces as possible. It then uses two sources of information - Rocketfuel's undns utility [47] and data from the Sarangworld project [40] - to determine the locations of these interfaces based on their DNS names. This step alone does not suffice for our purpose of clustering geographically co-located interfaces because: 1) several interfaces do not have a DNS name assigned to them, 2) rules for inferring the locations of all DNS names do not exist, and 3) incorrect locations are inferred for interfaces that have been misnamed. For IPs whose locations can be inferred from DNS names, the locations are validated by determining if they are consistent with the measured delays from traceroutes [28].

Third, to cluster interfaces for which a valid location was not determined, we develop an automated algorithm that clusters interfaces based on responses received from them when probed from a large number of vantage points. We probe all interfaces from all of iPlane's PlanetLab vantage points using ICMP ECHO probes. We use the TTL value in the response to estimate the number of hops on the reverse path back from every router to each of our vantage points. Our hypothesis is that routers in the same AS that are geographically nearby will have almost identical routing table entries and hence, take similar reverse paths back to each vantage point.

To translate this hypothesis into a clustering algorithm, each interface is associated with a reverse path length vector. This is a vector with as many components as the number of vantage points, and the $i^{th}$ component is the length of the reverse path from the interface back to the $i^{th}$ vantage point. We define the cluster distance between two vectors to be the L1 distance--the sum of the absolute differences between corresponding components, divided by the number of components. In our measurements, we have observed that the cluster distance between reverse path length vectors of co-located routers in an AS is normally less than $1$.

Based on the metric discussed above, we can now present a technique for assigning interfaces without known locations to clusters. We start by initializing our clusters to contain those interfaces for which a location has been determined. Interfaces that have been determined to be co-located in an AS are in the same cluster. For each cluster, we compute the median reverse path length vector, whose $i^{th}$ component is the median of the $i^{th}$ components of the vectors corresponding to all interfaces in the cluster. We then cluster all interfaces that do not belong to any cluster as follows. For each interface, we determine the cluster in the same AS as the interface, with whose median vector the interface's vector has the least cluster distance. If this minimum cluster distance is less than $1$, the interface is added to the chosen cluster, otherwise a new singleton cluster is created. This clustering algorithm, when executed on a typical traceroute output, clusters 762,701 interfaces into 54,530 clusters. 653,455 interfaces are in 10,713 clusters of size greater than $10$, while 21,217 interfaces are in singleton clusters.



2.3 Measuring the Internet Core


After clustering, iPlane can operate on a compact routing topology, where each node in the topology is a cluster of interfaces and each link connects two clusters. iPlane then seeks to determine a variety of link attributes that can be used to predict path performance. To achieve this goal, a centralized agent is used to distribute the measurement tasks such that each vantage point is assigned to repeatedly measure only a subset of the links. The centralized agent uses the compact routing topology to determine the assignments of measurement tasks to vantage points, communicates the assignment, and monitors the execution of the tasks. Only iPlane infrastructure nodes (namely, PlanetLab nodes) are used for these tasks.


2.3.1 Orchestrating the Measurement Tasks

There are three objectives to be satisfied in assigning measurement tasks to vantage points. First, we want to minimize the measurement load by measuring each link attribute from only a few vantage points (we employ more than one to correct for measurement noise). Second, the measurement should be load-balanced across all vantage points, i.e., each vantage point should perform a similar number of measurements. Third, in order to measure the properties of each link as accurately as possible, we measure every link in the topology from the vantage point that is closest to it.

We have developed a novel ``frontier'' algorithm to perform the assignment of tasks to vantage points. The algorithm works by growing a frontier rooted at each vantage point and having each vantage point measure only those links that are at its frontier. The centralized agent performs a Breadth-First-Search (BFS) over the measured topology in parallel from each of the vantage points. Whenever a vantage point is taken up for consideration, the algorithm performs a single step of the BFS by following one of the traceroute paths originating at the vantage point. If it encounters a link whose measurement task has been assigned already to another vantage point, it continues the BFS exploration until it finds a new link that has not been seen before. This process continues until all the link measurements have been assigned to some vantage point in the system.

The centralized agent uses the above algorithm to determine the assignment of tasks and then ships the tasklist to the respective vantage points. Each target link is identified by the traceroute path that the vantage point can use to reach the link and by its position within the traceroute path. If a vantage point is no longer capable of routing to the link due to route changes, the vantage point reports this to the centralized agent, which in turn reassigns the task to a different vantage point.

Most link attributes, however, cannot be directly determined by the vantage points. For instance, when measuring loss rates, a vantage point can only measure the loss rate associated with the entire path from the vantage point to the target link; the loss rates of individual links have to be inferred as a post-processing operation. Once all vantage points report their measurements back to the centralized agent, the agent can perform the BFS style exploration of the topology to infer link properties in the correct order. For instance, assume that a vantage point $v$ had probed the path $v,\ldots,x,y$ and obtained a (one-way) loss rate measurement of $l_{v,y}$ for the entire path. The centralized agent can then infer the loss rate along the link $(x,y)$ after inferring the loss rates for each of the links in $v,\ldots,x$, composing these individual loss rates to compute the loss rate $l_{v,x}$ along the segment $v
\ldots x$, and then calculating the loss rate for $(x,y)$ using the equation $(1 - l_{v,y}) = (1 - l_{v,x})\cdot(1 - l_{x,y})$. Since the link property inference is performed as a BFS traversal, we are guaranteed that loss rates for all the links along $v,\ldots,x$ have been inferred before we consider the link $(x,y)$.

In our current system, the centralized agent schedules and monitors roughly 2700K measurements per day, a management load that a single centralized agent can easily bear. Fault tolerance is an issue, but is addressed by a simple failover mechanism to a standby controller. Note that the processed data is served to applications from a replicated database to ensure high availability.


2.3.2 Measurement of Link Attributes

We next outline the details of the loss rate, bottleneck capacity and available bandwidth measurements performed from each vantage point. Previous research efforts have proposed specific ways to measure each of these properties; our goal is to integrate these techniques into a useful prediction system. Latencies of path segments can be derived directly from the traceroute data gathered while mapping the topology, and therefore do not need to be measured explicitly.

Loss Rate Measurements: We perform loss rate measurements along path segments from vantage points to routers in the core by sending out probes and determining the fraction of probes for which we get responses. We currently use the simple method of sending TTL-limited singleton ICMP probes with a 1000-byte payload. When the probe's TTL value expires at the target router, it responds with a ICMP error message, typically with a small payload. When a response is not received, one cannot determine whether the probe or the response was lost, but there is some evidence from previous studies that small packets are more likely to be preserved even when routers are congested [32]. We therefore currently attribute all of the packet loss to the forward path; the development of more accurate techniques is part of ongoing work.

Capacity Measurements: We perform capacity measurements using algorithms initially proposed by Bellovin [3] and Jacobson [24] that vary the packet size and determine the delay induced by increased packet sizes. For each packet size, a number of probes (typically 30-40) of that size are sent to an intermediate router and the minimum round-trip time is noted. The minimum round-trip time observed over many probes can be regarded as a baseline path latency measurement with minimal queueing delays. By performing this experiment for different packet sizes, one can determine the increased transmission cost per byte. When this experiment is performed for a sequence of network links in succession, the capacity of each link can be determined. Note that our capacity measurements may underestimate a cluster link if it consists of multiple parallel physical links.

Available Bandwidth Measurements: Once we have link capacities, we can probe for available bandwidth along path segments using packet dispersion techniques such as Spruce [50], IGI [21], Pathload [25]. A simple measurement is performed by sending a few, equally spaced, short probes at the believed bottleneck capacity of the path segment, and then measuring how much delay they induce. The slope of the delay increase will indicate how much background traffic arrived during the same time period as the probe. For instance, if the probes are generated with a gap of $\Delta_{in}$ through a path segment of capacity $C$ and if the measured gap between between the probe replies is $\Delta_{out}$, one can estimate the available bandwidth as $C \cdot (1 -
\frac{\Delta_{out} - \Delta_{in}}{\Delta_{in}})$. An important detail is that the packets have to be scheduled at the desired spacing, or else the measurement is not valid. Fortunately, even on heavily loaded PlanetLab nodes, it is possible to realize the desired scheduling most of the time.



2.4 Opportunistic Edge Measurements


To provide a comprehensive data set on which to infer current properties of paths to end-hosts, it is necessary for iPlane to maintain an up-to-date map of the network that extends to the very edge. However, the measurement techniques outlined above are unlikely to work as they, as most other active measurements, require end-hosts to respond to unsolicited ICMP, UDP or TCP packet probes. Also, measurements to end-hosts are frequently misinterpreted by intrusion detection systems as attacks. Hence, we pursue an opportunistic approach to data collection--measuring paths to end-hosts while interacting with them over normal connections. We participate in the popular file-distribution application BitTorrent [9] and gather measurements from our exchanges with the peers in this swarming system. Note that BitTorrent has the further desirable property that anyone can connect to anyone, allowing us to arrange measurements of multiple paths to participating edge hosts.

BitTorrent is used daily by thousands of end users to distribute large files. BitTorrent is one example of a large class of swarming data distribution tools. By participating in several BitTorrent swarms, we have the opportunity to interact with a large pool of end-hosts. We measure properties of the paths to peers while exchanging data with them as part of the swarming system.

We currently gather two kinds of measurements using our opportunistic measurement infrastructure.

$\bullet$
Packet traces of TCP flows to end-hosts. These traces provide information about packet inter-arrival times, loss rates, TCP retransmissions and round trip times. We use the inter-arrival times between data packets to measure bottleneck bandwidth capacities of paths from clients to vantage points, as described further in Section 3.
$\bullet$
Traceroutes to end-hosts. When a peer connects to our measurement node, we conduct a traceroute to that host. We record this data and add it to our atlas.


2.5 Performance Prediction


Next, we describe how to predict path properties between an arbitrary pair of nodes based on the above measurements. The prediction proceeds in two steps. First, we predict the forward and reverse paths connecting the two nodes. Second, we aggregate measured link-level properties to predict end-to-end path properties.

Figure 1: The path from $S$ to $D$ is obtained by composing a path from $S$ with a path going into $D$ from a vantage point close to $S$ ($V_1$). $BGP_1$ and $BGP_2$ are destinations in two random prefixes to which $S$ performs traceroutes.

Path Prediction We use a technique we earlier developed [31] based on composing observed path segments to predict unknown paths. Consider a source $S$ and destination $D$. If $S$ is a vantage point, then we simply return the measured path from $S$ to $D$. Else, we determine an appropriate intersection point $I$ in the measured subgraph of the Internet such that--(a) the AS hop count of the path $S.I.D$ is minimum, and (b) the latency from $S$ to the point where the path $S.I.D$ exits the first-hop AS is minimum, in that order (Figure 1). The underlying principle is similarity of routes, i.e., with a sufficiently large number of vantage points, the path to a destination ($D$) from any node ($S$) will be similar to the path from a vantage point or router ($I$) located nearby. Condition (a) encodes the default path selection criterion used by BGP in the absence of conflicting local preference policies. Condition (b) encodes the default early exit intradomain routing policy. Note that the above technique is guaranteed to return a path (albeit an inflated one), since every path of the form $S.V.D$, for each vantage point $V$, belongs to the measured subgraph.

As we noted earlier, we make measurements to BGP atoms rather than to all destinations. In [31], we note that adding a small number of measured paths originating from the client $S$ significantly improves the prediction accuracy for paths sourced at $S$. Using these measurements, the path from $S$ to $D$ is $S$ to $I$ to $D$'s atom to $D$. If there is a measurement of the last hop from $D$'s atom to $D$, we use it; otherwise, we estimate it using a representative node in the atom (e.g., from BitTorrent measurements). Briefly summarizing the main results from [31], we can predict the AS path exactly right for around $70\%$ of the paths evaluated, and the latency estimates obtained using this model were significantly better than those yielded by Vivaldi [11], a popular network coordinate system.

Path Properties Given predicted paths as above, we can estimate end-to-end properties by aggregating link-level properties. For example, we predict TCP transfer time using widely accepted models [35,5]. For this, we separately predict the forward and reverse paths between the source and the destination. The latency on the forward path $S.I.D$-$atom.D$ is estimated as the sum of the latency estimates for each segment. We similarly estimate the latency along the reverse path, and then compute the RTT between the two end-hosts to be the sum of our latency estimates along the forward and reverse paths. The loss rate on the predicted forward path is estimated from the probability of a loss on any of its constituent links while bandwidth is the minimum value across the links. The access link capacities of these end-hosts, if available based on BitTorrent measurements to hosts in the same $/24$ prefixes, are also used to estimate the end-to-end bottleneck bandwidth.

Recently, He et al. [20] argued that the best way to accurately predict TCP throughput is to send TCP flows and use history-based predictors. Although we have not implemented these, our use of passive BitTorrent logs is amenable to incorporating such predictors.

2.6 Securing iPlane

iPlane allows untrusted users to contribute measurements, so it is vulnerable to attacks aimed at polluting its information. For instance, a client can claim to have better connectivity than actuality in order to improve its position within an overlay service that uses iPlane. iPlane reduces this risk by using client data only for those queries issued by the same client; falsified measurements will not affect the queries issued by other clients.

We do however trust traceroute servers to provide unbiased data, though the traceroute servers are not under our control. An ISP hosting a traceroute server might bias its replies from the server to better position its clients, for example, to attract more BitTorrent traffic and thereby generate more revenue. We have the ability to use verification to address this - compare the results from multiple vantage points for consistency - but have not implemented it yet.


2.7 Query Interface

The query interface exported by iPlane must be carefully designed to enable a diverse range of applications. Our current implementation of the query interface exposes a database-like view of path properties between every pair of end-hosts in the Internet. For every source-destination pair, there exists a row in the view with iPlane's predicted path between these hosts and the predicted latency, loss rate, and available bandwidth along this path. Any query to iPlane involves an SQL-like query on this view - selecting some rows and columns, joining the view with itself, sorting rows based on values in certain columns, and so on. The database view is merely an abstraction. iPlane does not compute apriori the entire table comprising predictions for every source-destination pair; instead it derives necessary table entries on-demand.

For example, a CDN can determine the closest replica to a given client by selecting those rows that predict the performance between the client and any of the CDN's replicas. A suitable replica can then be determined by sorting these rows based on a desired performance metric. To choose a good detour node for two end-hosts to conduct VoIP, the rows predicting performance from the given source can be joined with the set of rows predicting performance for the given destination. A good detour is one that occurs as the destination in the first view and as the source in the second view, such that the composed performance metrics from these rows is the best. These queries can be invoked in any one of the following ways.

Download the Internet Map: We have implemented a library that provides an interface to download the current snapshot of the entire annotated Internet map or a geographic region, to process the annotated map, and to export the above SQL-like view. An application simply links against and invokes the library locally.

On-the-fly Queries: Applications that do not wish to incur the costs of downloading the annotated map and keeping it up-to-date, can query a remote iPlane service node using non-local RPCs. Note that clients of CDNs, such as Akamai and Coral, typically tolerate some indirection overhead in determining the nearest replica. To support such applications, iPlane downloads the annotated map of the Internet to every PlanetLab site, and then provides an RPC interface to the data. Further, as some applications might need to make multiple back-to-back queries to process iPlane's measurements, we assist the application in lowering its overheads by allowing it to upload a script that can make multiple local invocations of iPlane's library. The current implementation requires that this script be written in Ruby, as Ruby scripts can be executed in a sandboxed environment and with bounded resources [49]. The output of the script's execution is returned as the response to the RPC.

Network Newspaper: Apart from downloading the Internet graph and issuing on-the-fly queries, a third model that we plan to support is a publish-subscribe interface that allows users to register for information updates about specific portions of the Internet graph. This interface allows users to subscribe to their ``view'' of the Internet, i.e., all paths originating from a user to all BGP atoms, or insert triggers to be notified of specific events, e.g., when a critical link fails. Implementing such an interface is part of our future work.

Figure 2: Overall architecture of iPlane.

The various components in our current implementation of iPlane, and the interaction between these components is depicted in Figure 2.


3 System Setup and Evaluation

In this section, we present details of our deployment of iPlane. We provide an overview of the measurements we conducted as part of our deployment. We also outline the tests we conducted to validate our measurements. All of our validation is performed on paths between PlanetLab nodes; our goal in the future is to use client measurements (e.g., DIMES [43]) to broaden the validation set. The number of PlanetLab nodes used varies with each experiment because of the variable availability of some nodes.

3.1 Measuring the Core

Figure 3: Distribution of errors in (a) latency, and (b) loss rate estimation.
(a)
(b)

We first consider results from a typical run of our mapping process. We performed traceroutes from PlanetLab nodes in $163$ distinct sites. The targets for our trace-routes were $.1$ addresses in each of 91,498 prefixes determined from the RouteViews BGP snapshot, though measuring paths to one address in each BGP atom should suffice. We probed all interfaces observed in our measured topology with UDP and ICMP probes, and clustered the interfaces based on their responses.

Once a map of the Internet's core was gathered, we employed our ``frontier'' BFS algorithm to determine paths to be probed from each of the $385$ PlanetLab nodes present at the $163$ sites used; for link metrics, we use multiple nodes per site. To determine the properties of 270,314 inter-cluster links seen in our measured topology, each vantage point was assigned to measure only around 700 paths. Loss rate, capacity, and available bandwidth were measured for each of the assigned paths. These measurements were then processed to determine properties for every cluster-level link in our measured topology.

To validate the predictive accuracy of iPlane, we compared properties of paths between PlanetLab nodes with the corresponding values predicted by iPlane. We measured the latency and loss rate along every path beween any two PlanetLab nodes. To predict the performance, we assume that we have the probe information collected by the other $161$ sites, excluding the source and destination under consideration. We then added 10 traceroutes from the source and destination to random nodes to simulate the behavior of participating clients. Each experiment was performed independently to ensure no mixing of the measurement and validation set. Figure 3 compares the latency and loss rate estimates made by iPlane with the true values. For $77\%$ of paths, iPlane's latency estimates have error less than $20$ms, and for $82\%$ of paths, loss rate estimates have error less than $10\%$.

Figure 4: Rank correlation coefficient between actual and predicted TCP throughput.

Further, we evaluated how predictive of path performance are iPlane's estimates of latency and loss rate in combination. The desired property of these estimates is that they help distinguish between paths with good and bad performance. We compared the order of paths from each PlanetLab node in terms of actual and predicted performance. For each node, we ranked all other nodes in terms of TCP throughput, considering throughput to be inversely proportional to latency and the square root of loss rate [35]. These rankings were computed independently using measured path properties and using iPlane's predictions for these properties. Figure 4 plots the correlation coefficient between the actual and iPlane predicted rankings across all PlanetLab nodes. For $80\%$ of the nodes, the correlation coefficient is greater than $0.7$.


3.2 Scalability


Table 2: Complexity of measurements techniques used in iPlane based on the following assumptions. A UDP/ICMP probe is 40 bytes. A traceroute incurs a total of 500B on average. The per-link loss rate, available bandwidth, and capacity measurements require 200KB, 100KB, and 200KB of probe traffic respectively.
Measurement Task Tool / Technique Frequency Probing rate / node

Topology Mapping

traceroute

Once a day

$200$ vantage points $\times$ $50$K atoms -- $2.5$Kbps

Clustering

UDP probes for source-address-based alias resolution, ICMP-ECHO probes for RTTs and reverse TTLs

One day
every week

$100$ vantage points $\times$ $800$K interfaces -- $6$Kbps

Capacity measurements

``frontier" algorithm applied to cluster-level topology for path assignment,
pathchar for bandwidth capacity

Once a day

$400$ vantage points $\times$ $700$ links -- $13$Kbps

Loss rate and available bandwidth measurements

``frontier" algorithm for path assignment, TTL-limited probes for loss rate,
spruce for available bandwidth

Continuous
(every $6$ hours)

$400$ vantage points $\times$ $700$ links -- $80$Kbps



We now discuss the measurement load required to generate and maintain a frequently refreshed map of the Internet. The measurement tasks performed by iPlane have two primary objectives--mapping of the Internet's cluster-level topology and determination of the properties of each link in the measured topology. Measurement of link properties incurs higher measurement overhead when compared to the probe traffic needed to perform a traceroute, but scales better. With more vantage points, the topology discovery traffic per node remains the same, but the overhead per node for measuring link metrics scales down, allowing the same fidelity for less overhead per node. The measurement load associated with each technique in our measurement apparatus is summarized in Table 2. These numbers assume the availability of 400 PlanetLab nodes at 200 sites. Our main result is that iPlane can produce an updated map of the Internet's routing topology every day with as little as $10$Kbps of probe traffic per vantage point, and update the map of link-level attributes once every 6 hours with around $100$Kbps of probe traffic per vantage point, suggesting that iPlane can refresh the Internet map frequently.

3.3 Stationarity of Measurements

Figure 5: Stationarity of measurements over different intervals over the course of a day.
(a)
(b)

iPlane's measurements change over time with changes in the routes in the Internet and the traffic they carry. We again use PlanetLab data to estimate whether it suffices for iPlane to update its map every 6 hours. We are currently in the process of evaluating the stationarity of path properties for non-PlanetLab destinations as well.

Over a period of 2 days, we measured the latency and loss rate between PlanetLab nodes once every 30 minutes. For this study, we used a dataset of 174 PlanetLab sites spanning 29 countries. In every interval, we computed for each node the ranking of all other nodes in terms of TCP throughput. To evaluate the flux in path properties over a 30 minute timescale, we compared these rankings between adjacent 30 minute intervals. For each PlanetLab node, we computed the correlation coefficient between the ranking vectors from adjacent intervals as well as computed the intersection between the top 10 nodes in these ranking vectors. To compare this with the flux in measurements over longer timescales, we also performed these computations across intervals 1 hour, 2 hours, 4 hours, 8 hours, 16 hours and 24 hours apart.

Figure 5(a) shows that the median correlation coefficient between the rankings is greater than $0.8$ across all intervals from 30 minutes to a day. Similarly, Figure 5(b) shows that in the median case, 7 of the top 10 nodes in this ranking are identical on timescales from 30 minutes to a day. Though these results are only for paths between PlanetLab nodes, they seem to indicate that there is little value in updating the map more frequently than once every few hours, compared to once every 30 minutes.

3.4 Measurements to End-Hosts

To measure the edges of the Internet, we deployed a modified BitTorrent client on $367$ PlanetLab nodes. As described in Section 2.4, our infrastructure for measuring the edge involves the millions of users who frequently participate in the BitTorrent filesharing application. Every hour, we crawl well-known public websites that provide links to several thousand .torrent files to put together a list of $120$ popular swarms. The number of swarms for consideration was chosen so as to ensure the participation of several of our measurement vantage points in each swarm. The number of PlanetLab nodes designated to a swarm is proportional to the number of peers participating in it.

Each PlanetLab node runs a BitTorrent client that we have modified in several ways to aid in our measurements. First, the modified client does not upload any data nor does it write any data that it downloads onto disk. Second, our client severs connections once we have exchanged $1$MB of data, which suffices for purposes of our measurements. Finally, we introduce a shadow tracker--a database that coordinates measurements among all PlanetLab nodes participating in a single swarm. Instead of operating only on the set of peers returned by the original tracker for the swarm, our modified client also makes use of the set of peers returned to any measurement node. Clients preferentially attempt to connect and download data from peers that have not yet been measured by a sufficient number of vantage points. These modifications are crucial for measurement efficiency and diversity since typical BitTorrent trackers permit requesting only a restricted set ($50$-$100$) of participating peers once every $30$ minutes or more. Such short lists are quickly exhausted by our modified client.

During a $48$ hour period, our measurement nodes connected to 301,595 distinct IP addresses, and downloaded sufficient data to measure the upload bandwidth capacity from 70,428. These hosts span 3591 distinct ASs, 19,639 distinct BGP prefixes, and 160 different countries.

3.5 Validation of BitTorrent capacity measurements

Figure 6: CDFs of estimated bandwidth capacity on paths between PlanetLab nodes as measured by iPlane and $S^{3}$.

Our edge bandwidth capacity measurement relies on inter-arrival times observed between data packets in the connections we maintain with BitTorrent peers. We implemented the multiQ [27] technique to infer end-to-end bottleneck bandwidth capacity from these inter-arrival times. Although the accuracy of multiQ presented in previous studies is encouraging, the unique properties of PlanetLab motivated us to provide further validation. To verify that multiQ yields reasonable data with short TCP traces in the presence of cross traffic on machines under heavy load, we compared our measurements with those made by $S^{3}$ [13].

We setup a test torrent and had our measurement clients running on $357$ PlanetLab nodes participate in this torrent. From this setup, we opportunistically measured the bottleneck bandwidth capacities between these PlanetLab nodes. The dataset we gathered from this experiment had 10,879 paths in common with measurements made by $S^{3}$ on the same day. Figure 6 compares the bandwidth capacities measured by the two methods. The measurements made by iPlane closely match those of $S^{3}$ for capacities less than $10$ Mbps. At higher bandwidth capacities, they are only roughly correlated. We attribute this difference to the use of user-level timestamps by $S^{3}$. As inter-packet spacing can be rather small for high capacity paths, user-level timestamps are likely to be inaccurate in the highly loaded PlanetLab environment. Our measurement setup makes use of kernel-level timestamps and is therefore less sensitive to high CPU load. For typical access link bandwidths, the two tools produce similar data; the value of using BitTorrent is that it works with unmodified clients that sit behind firewalls or NATs that would drop active measurement probes. The more discernible steps in the iPlane line in Figure 6 are at 10Mbps, 45Mbps (T3), and 100Mbps, which correspond to typical ISP bandwidth classes.

3.6 Clustering of end-hosts

Figure 7: CDF of the ratio of maximum to minimum measured bandwidth capacity for $/24$ address prefixes with multiple measurements from the same vantage point across time.

Although the data produced by our opportunistic strategy is extensive, it is by no means complete. Not every client participates in popular torrents. In Figure 7, we explore the validity of using BitTorrent measurements to predict the performance of other clients in the same prefix. For every $/24$ prefix in which we have measurements to multiple end-hosts from the same vantage point, we compute the ratio of the maximum to the minimum measured bandwidth capacity. For $70\%$ of $/24$ prefixes, the capacities measured differ by less than $20\%$.

4 Application Case Studies

In this section, we show how applications can benefit from using iPlane. We evaluate three distributed services for potential performance benefits from using iPlane.

4.1 Content Distribution Network

Figure 8: CDF of download times from replicas in the CDN chosen by the iPlane and from replicas closest in terms of latency. Each download time is the median of $5$ measurements.

Content distribution networks (CDNs) such as Akamai, CoDeeN and Coral [1,52,16] redirect clients to a nearby replica. The underlying assumption is that distance determines network performance. However, there is more to network performance than just distance, or round trip time. TCP throughput, for example, depends on both distance and loss rate [35,5]. Even for small web documents, loss of a SYN or a packet during slow start can markedly inflate transfer time. A CDN using iPlane can track the RTT, loss rate, and bottleneck capacity from each replica to the rest of the Internet. The CDN can then arrange for its name servers to redirect the client to optimize using the model of its choice.

We emulate a small CDN comprising 30 randomly chosen PlanetLab nodes. Each node serves 3 files of sizes 10KB, 100KB and 1MB. We use 141 other PlanetLab nodes to emulate clients. Each client downloads all 3 files from the replica that provides the best TCP throughput as predicted by the PFTK model [35] using iPlane's estimates of RTT and loss rate, and from the replica closest in terms of actual measured RTT. Note that this comparison is against an optimum that cannot be achieved without extensive probing. A real CDN will only have estimated RTTs available. Figure 8 compares the download times experienced by the clients in either case, excluding the latency of redirecting to the replica. Choosing the replica for optimized TCP throughput based on iPlane's predictions provides slightly better performance than choosing the closest replica. Though these results are only indicative, they suggest that iPlane with its ability to provide multi-attribute network performance data will be more effective than systems such as OASIS [17] that simply optimize for RTT.

4.2 BitTorrent

Figure 9: CDFs of BitTorrent download completion times with and without informed peer selection at the tracker.


We next show how iPlane can enable informed peer selection in popular swarming systems like BitTorrent. In current implementations, a centralized BitTorrent tracker serves each client a random list of peers. Each client enforces a tit-for-tat bandwidth reciprocity mechanism that incents users to contribute more upload bandwidth to obtain faster downloads. However, the same mechanism also serves to optimize path selection at a local level--peers simply try uploading to many random peers and eventually settle on a set that maximizes their download rate. Because reasoning about peer quality occurs locally at each client, each client needs to keep a large pool of directly connected peers ($60$-$100$ for typical swarms) even though at any time only a few of these ($10$-$20$) are actively engaged in data transfer with the client. This overhead and consequent delayed convergence is fundamental: with only local information, peers cannot reason about the value of neighbors without actively exchanging data with them. iPlane's predictions can overcome the lack of prior information regarding peer performance and can thus enable a clean separation of the path selection policy from the incentive mechanism.

We built a modified tracker that uses iPlane for informed peer selection. Instead of returning random peers, the tracker uses the iPlane's loss rate and latency estimates to infer TCP throughput. It then returns a set of peers, half of which have high predicted throughput and the rest randomly selected. The random subset is included to prevent the overlay from becoming disconnected (e.g., no US node preferring a peer in Asia).

We used our modified tracker to coordinate the distribution of a $50$ megabyte file over $150$ PlanetLab nodes. We measured the time taken by each of the peers to download the file after the seed was started. Figure 9 compares the download times observed with iPlane predictions against those of peerings induced by Vivaldi coordinates [11] and an unmodified tracker. Informed peer selection causes roughly $50\%$ of peers to have significantly lower download times.

Although preliminary, these performance numbers are encouraging. We believe that better use of information from the iPlane can lead to even further improvements in performance. Our selection of $50\%$ as the fraction of random peers was arbitrary, and we are currently investigating the tradeoff between robustness and performance, as well as the degree to which these results extend to swarms with a more typical distribution of bandwidths.

4.3 Voice Over IP

Figure 10: Comparison of (a) loss rate and (b) jitter with and without use of iPlane for end-to-end VoIP paths.

(a)

(b)

Voice over IP (VoIP) is a rapidly growing application that requires paths with low latency, loss and jitter for good performance. Several VoIP implementations such as Skype [45] require relay nodes to connect end-hosts behind NATs/firewalls. Choosing the right relay node is crucial to providing acceptable user-perceived performance [39]. Reducing end-to-end latency is important since humans are sensitive to delays above a threshold. Low loss rates improve sound quality and reduce throughput consumed by compensating codecs. Measures of user-perceived sound quality such as mean opinion score [51] have been shown to be highly correlated with loss rate and end-to-end delay. Thus, VoIP applications can benefit from iPlane's predictions of latency and loss rate in choosing the best possible relay node.

To evaluate iPlane's ability to successfully pick good relay nodes, we emulated VoIP traffic patterns on PlanetLab. We considered $384$ pairs of PlanetLab nodes, chosen at random, as being representative of end-hosts participating in a VoIP call. Between each pair, we emulated a call by sending a $10$KBps UDP packet stream via another PlanetLab node chosen as the relay node. We tried $4$ different relay options for each pair chosen based on (i) the iPlane's estimates of latency and loss rate, (ii) latency to the source, (iii) latency to the destination, and (iv) random choice. The iPlane-informed choice was obtained by first querying for the $10$ relay options that minimize end-to-end loss and then, choosing the one that minimized end-to-end delay among these options.

Each emulated call lasted for $60$ seconds, and the end-to-end loss rate and latency were measured. Figure 10(a) shows that significantly lower loss rates were observed along relay paths chosen based on iPlane's predictions. Additionally, Figure 10(b) shows that iPlane also helps to reduce jitter, which we computed as the standard deviation of end-to-end latency. These results demonstrate the potential for the use of iPlane in VoIP applications.

5 Related Work

iPlane bridges and builds upon ideas from network measurement, performance modeling, Internet tomography, and recent efforts towards building a knowledge plane for the Internet. We believe that an Internet-scale instantiation of iPlane is greater than the sum of its parts, and relate individual contributions to prior work.

Information Plane Clark et al. [8] pioneered the broad vision of a knowledge plane to build large-scale, self-managing and self-diagnosing networks based on tools from AI and cognitive science. Several research efforts have since addressed pieces of this problem.

Several efforts have looked at monitoring end-host performance and in optimizing the query processing engine of the information plane. Examples include Sophia [53], PIER [23], and IrisNet [18]. The above systems have a different focus than ours. They manage information about nodes (e.g., PlanetLab nodes, routers in an ISP, or sensors) under control of the information plane. We target predictions of path performance at Internet-scale.

Link Metrics IDMaps [15] is an early example of a network information service that estimates the latency between an arbitrary pair of nodes using a small set of vantage points as landmarks. Subsequently, Ng and Zhang [34] discovered that Internet distances can be embedded on to a low-dimensional Euclidean space. Such embeddings can be used to predict latencies between a large number of nodes by measuring latencies from a small number of vantage points to these nodes--a methodology refined by several others [54,10,11,44]. A key limitation of these techniques is that they treat the Internet as a black box and are only predictive, i.e., they do not explain why, if at all, their predictions are correct. As a result, they have serious systematic deficiencies, e.g., a significant fraction of Internet paths are known to have detours [41], however, metric embeddings obey the triangle inequality and will predict no detours.

Our previous work on a structural technique [31] to predict Internet paths and latencies, and experiences reported by independent research groups with respect to latency prediction [17], available bandwidth estimation [22], and the practical utility of embedding techniques [30] echo the need for structural approaches to predict sophisticated path metrics.

Inference Techniques Chen et al. [7] proposed an algebraic approach to infer loss rates on paths between all pairs of nodes based on measured loss rates on a subset of the paths. Duffield et al. [14] proposed a multicast-based approach to infer link loss rates by observing loss correlations between receivers. Rocketfuel [47] estimates ISP topologies by performing traceroutes from a set of vantage points, while the Doubletree [12] system efficiently prunes redundant traceroutes. Our frontier (Section 2.3) algorithm to efficiently target specific links for measurement is similar in spirit.

Passive Measurements Padmanabhan et al. [36] and Seshan et al.[42] propose passive measurements at Web servers and end-hosts respectively to predict path metrics. PlanetSeer by Zhang et al. [55] is a failure monitoring system that uses passive measurements at CDN caches under their control to diagnose path failures postmortem. Jaiswal et al. [26] propose a ``measurements-in-the-middle'' technique to infer end-to-end path properties using passive measurements conducted at a router. In contrast to these systems that perform passive measurements of existing connections, we participate in BitTorrent swarms and opportunistically create connections to existing peers for the explicit purpose of observing network behavior. Previously, opportunistic measurements have relied on sprurious traffic in the Internet [6]. iPlane could also validate and incorporate measurement data from passive measurement sources, such as widely deployed CDNs, and such integration is part of future work.

6 Conclusion


The performance and robustness of overlay services critically depends on the choice of end-to-end paths used as overlay links. Today, overlay services face a tension between minimizing redundant probe overhead and selecting good overlay links. More importantly, they lack an accurate methods to infer path properties between an arbitrary pair of end-hosts. In this paper, we showed that it is possible to accurately infer sophisticated path properties between an arbitrary pair of nodes using a small number of vantage points and existing infrastructure. The key insight is to systematically exploit the Internet's structural properties. Based on this observation, we built the iPlane service and showed that it is feasible to infer a richly annotated link-level map of the Internet's routing topology once every few hours. Our case studies suggest that iPlane can serve as a common information plane for a wide range of distributed services such as content distribution, file swarming, and VoIP.

Acknowledgments

We would like to thank Jay Lepreau, Ratul Mahajan, KyoungSoo Park, Rob Ricci, Neil Spring and the anonymous OSDI reviewers for their valuable feedback on earlier versions of this paper. We also thank Peter Druschel for serving as our shepherd. This research was partially supported by the National Science Foundation under Grants CNS-0435065 and CNS-0519696.

Bibliography

1
Akamai, Inc. home page.
https://www.akamai.com.

2
D. G. Andersen, H. Balakrishnan, M. F. Kaashoek, and R. Morris.
Resilient Overlay Networks.
In SOSP, 2001.

3
S. Bellovin.
A best-case network performance model.
Technical report, ATT Research, 1992.

4
A. Broido and kc claffy.
Analysis of routeViews BGP data: policy atoms.
In Network Resource Data Management Workshop, 2001.

5
N. Cardwell, S. Savage, and T. Anderson.
Modeling TCP latency.
In INFOCOM, 2000.

6
M. Casado, T. Garfinkel, W. Cui, V. Paxson, and S. Savage.
Opportunistic measurement: Extracting insight from spurious traffic.
In HotNets, 2005.

7
Y. Chen, D. Bindel, H. Song, and R. H. Katz.
An algebraic approach to practical and scalable overlay network monitoring.
In SIGCOMM, 2004.

8
D. D. Clark, C. Partridge, J. C. Ramming, and J. T. Wroclawski.
A knowledge plane for the Internet.
In SIGCOMM, 2003.

9
B. Cohen.
Incentives build robustness in BitTorrent.
In P2PEcon, 2003.

10
M. Costa, M. Castro, A. Rowstron, and P. Key.
PIC: Practical Internet coordinates for distance estimation.
In ICDCS, 2004.

11
F. Dabek, R. Cox, F. Kaashoek, and R. Morris.
Vivaldi: A decentralized network coordinate system.
In SIGCOMM, 2004.

12
B. Donnet, P. Raoult, T. Friedman, and M. Crovella.
Efficient algorithms for large-scale topology discovery.
In SIGMETRICS, 2005.

13
C. Dovrolis, P. Ramanathan, and D. Moore.
Packet dispersion techniques and a capacity estimation methodology.
IEEE/ACM Transactions on Networking, 2004.

14
N. G. Duffield, F. L. Presti, V. Paxson, and D. F. Towsley.
Inferring link loss using striped unicast probes.
In INFOCOM, 2001.

15
P. Francis, S. Jamin, C. Jin, Y. Jin, D. Raz, Y. Shavitt, and L. Zhang.
IDMaps: A global Internet host distance estimation service.
IEEE/ACM Transactions on Networking, 2001.

16
M. J. Freedman, E. Freudenthal, and D. Mazières.
Democratizing content publication with Coral.
In NSDI, 2004.

17
M. J. Freedman, K. Lakshminarayanan, and D. Mazieres.
OASIS: Anycast for any service.
In NSDI, 2006.

18
P. Gibbons, B. Karp, Y. Ke, S. Nath, and S. Seshan.
IrisNet: An architecture for a world-wide sensor web.
IEEE Pervasive Computing, 2(4), 2003.

19
R. Govindan and H. Tangmunarunkit.
Heuristics for Internet map discovery.
In INFOCOM, 2000.

20
Q. He, C. Dovrolis, and M. Ammar.
On the predictability of large transfer TCP throughput.
In SIGCOMM, 2005.

21
N. Hu and P. Steenkiste.
Evaluation and characterization of available bandwidth probing techniques.
IEEE JSAC, 21(6), 2003.

22
N. Hu and P. Steenkiste.
Exploiting Internet route sharing for large scale available bandwidth estimation.
In IMC, 2005.

23
R. Huebsch, J. M. Hellerstein, N. L. Boon, T. Loo, S. Shenker, and I. Stoica.
Querying the Internet with PIER.
In VLDB, 2003.

24
V. Jacobson.
Pathchar.
ftp://ftp.ee.lbl.gov/pathchar.

25
M. Jain and C. Dovrolis.
End-to-end available bandwidth: measurement methodology, dynamics, and relation with TCP throughput.
In SIGCOMM, 2002.

26
S. Jaiswal, G. Iannaccone, C. Diot, J. Kurose, and D. Towsley.
Formal analysis of passive measurement inference techniques.
In INFOCOM, 2006.

27
S. Katti, D. Katabi, C. Blake, E. Kohler, and J. Strauss.
MultiQ: Automated detection of multiple bottleneck capacities along a path.
In IMC, 2004.

28
E. Katz-Bassett, J. P. John, A. Krishnamurthy, D. Wetherall, T. Anderson, and Y. Chawathe.
Towards IP geolocation using delay and topology measurements.
In IMC, 2006.

29
S. Lee, P. Sharma, S. Banerjee, S. Basu, and R. Fonseca.
Measuring bandwidth between planetlab nodes.
In PAM, 2005.

30
E. K. Lua, T. Griffin, M. Pias, H. Zheng, and J. Crowcroft.
On the accuracy of embeddings for Internet coordinate systems.
In IMC, 2005.

31
H. V. Madhyastha, T. E. Anderson, A. Krishnamurthy, N. Spring, and A. Venkataramani.
A structural approach to latency prediction.
In IMC, 2006.

32
R. Mahajan, N. Spring, D. Wetherall, and T. Anderson.
User-level Internet path diagnosis.
In SOSP, 2003.

33
D. Meyer.
RouteViews.
https://www.routeviews.org.

34
T. S. E. Ng and H. Zhang.
Predicting Internet network distance with coordinates-based approaches.
In INFOCOM, 2002.

35
J. Padhye, V. Firoiu, D. Towsley, and J. Kurose.
Modeling TCP throughput: A simple model and its empirical validation.
CCR, 28(4), 1998.

36
V. N. Padmanabhan, L. Qiu, and H. J. Wang.
Passive network tomography using bayesian inference.
In IMW, 2002.

37
A. Parker.
CacheLogic.
https://www.cachelogic.com/research/slide1.php.

38
L. Peterson.
Personal communication.

39
S. Ren, L. Guo, and X. Zhang.
ASAP: an AS-aware peer-relay protocol for high quality VoIP.
In ICDCS, 2006.

40
Sarangworld project.
https://www.sarangworld.com/TRACEROUTE/.

41
S. Savage, T. Anderson, A. Aggarwal, D. Becker, N. Cardwell, A. Collins, E. Hoffman, J. Snell, A. Vahdat, G. Voelker, and J. Zahorjan.
Detour: a case for informed Internet routing and transport.
IEEE Micro, 19(1), 1999.

42
S. Seshan, M. Stemm, and R. Katz.
SPAND: Shared passive network performance discovery.
In USITS, 1997.

43
Y. Shavitt and E. Shir.
DIMES: Let the Internet measure itself.
CCR, 35(5), 2005.

44
Y. Shavitt and T. Tankel.
On the curvature of the Internet and its usage for overlay construction and distance estimation.
In INFOCOM, 2004.

45
Skype home page.
https://www.skype.com.

46
N. Spring, R. Mahajan, and T. Anderson.
Quantifying the causes of path inflation.
In SIGCOMM, 2003.

47
N. Spring, R. Mahajan, D. Wetherall, and T. Anderson.
Measuring ISP topologies with Rocketfuel.
IEEE/ACM Transactions on Networking, 2004.

48
N. Spring, D. Wetherall, and T. Anderson.
Reverse-engineering the Internet.
In HotNets, 2003.

49
N. Spring, D. Wetherall, and T. Anderson.
Scriptroute: A public Internet measurement facility.
In USITS, 2003.

50
J. Strauss, D. Katabi, and F. Kaashoek.
A measurement study of available bandwidth estimation tools.
In IMC, 2003.

51
S. Tao, K. Xu, A. Estepa, T. Fei, L. Gao, R. Guerin, J. Kurose, D. Towsley, and Z.-L. Zhang.
Improving VoIP quality through path switching.
In INFOCOM, 2005.

52
L. Wang, K. Park, R. Pang, V. S. Pai, and L. L. Peterson.
Reliability and security in the CoDeeN content distribution network.
In USENIX, 2004.

53
M. Wawrzoniak, L. Peterson, and T. Roscoe.
Sophia: An information plane for networked systems.
In HotNets-II, 2003.

54
B. Wong, A. Slivkins, and E. G. Sirer.
Meridian: A lightweight network location service without virtual.
In SIGCOMM, 2005.

55
M. Zhang, C. Zhang, V. Pai, L. Peterson, and R. Wang.
PlanetSeer: Internet path failure monitoring and characterization in wide-area services.
In OSDI, 2004.

56
H. Zheng, E. K. Lua, M. Pias, and T. Griffin.
Internet routing policies and round-trip-times.
In PAM, 2005.

About this document ...

iPlane: An Information Plane for Distributed Services

This document was generated using the LaTeX2HTML translator Version 2002-2-1 (1.71)

Copyright © 1993, 1994, 1995, 1996, Nikos Drakos, Computer Based Learning Unit, University of Leeds.
Copyright © 1997, 1998, 1999, Ross Moore, Mathematics Department, Macquarie University, Sydney.

The command line arguments were:
latex2html -split 0 -show_section_numbers -local_icons iplane.tex

The translation was initiated by Harsha Madhyastha on 2006-09-06


next_inactive up previous
Harsha Madhyastha 2006-09-06