Ramesh Govindan and Anoop Reddy. An analysis of internet inter-domain topology and route stability. In INFOCOM (2), pages 850--857, 1997. 41  Home/Search   Document Details and Download   Summary   Related Articles   Check

....the Global Positioning System (GPS) as a means for improving accuracy of active probes [21] That infrastructure is quite similar to Surveyor [11] which was used to gather data used in our study. Internet topology and routing characteristics have also been widely studied. Examples include [22] [23], 24] These studies inform our work with respect to the structural characteristics of end to end Internet paths. A variety of methods have been employed to model network packet traffic including queuing and auto regressive techniques [25] While these models can be parameterized to recreate ....

R. Govindan and A. Reddy, "An analysis of internet inter-domain topology and route stability," in Proceedings of IEEE INFOCOM '97, Kobe, Japan, April 1997. 12

....shed considerable light on factors including performance, stability and growth. Examples of these that relate to our work include studies that investigate basic characteristics of packet dynamics [17] 18] 19] 20] 21] and studies that assess Internet routing and path characteristics [22] [23], 24] Of these, the work that is perhaps most similar to ours is that of Zhang et al. 21] In that study, the authors use a large measurement data set to define characteristics of network path constancy related to packet loss, packet delay and TCP throughput. We do not specifically treat ....

R. Govindan and A. Reddy, "An analysis of internet inter-domain topology and route stability," in Proceedings of IEEE INFOCOM '97, Kobe, Japan, April 1997.

....that the majority of BGP messages consisted of redundant pathological announcements. 4] further identified the origins of certain pathological behavior. They also showed that routing instability had been significantly reduced in the core network by software improvements. Govindan and Reddy [2] studied the Internet topology and routing stability several years ago. They found that routes to prefixes were highly available and stable at that time, but the mean reachability duration for a prefix decreases with the Internet growth. The Internet has grown rapidly since this study and more ....

R. Govindan and A. Reddy. An analysis of internet interdomain topology and route stability. In INFOCOM, pages 850--857, 1997.

....knowledge about the Internet structure, the research on the subject has recently produced many contributions. More speci cally, there is a wide research area focusing on the discovery of the topology underlying the Internet structure, either at the AS and at the router level (see, for example, [10, 11, 18]) Other researchers concentrate more directly on the above mentioned relationships and on the hierarchy that they induce on the set of ASes. Govindan and Reddy [10] study the interplay between the degree of the ASes and their rank in the hierarchy, where the degree of an AS is the number of ASes ....

....the discovery of the topology underlying the Internet structure, either at the AS and at the router level (see, for example, 10, 11, 18] Other researchers concentrate more directly on the above mentioned relationships and on the hierarchy that they induce on the set of ASes. Govindan and Reddy [10] study the interplay between the degree of the ASes and their rank in the hierarchy, where the degree of an AS is the number of ASes that have some kind of relationship with it. Gao [6] studies, for the rst time, the following problem. ASes are the vertices of a graph (AS graph) where two ASes ....

R. Govindan and A. Reddy. An analysis of internet inter-domain topology and route stability. In Proc. IEEE INFOCOM 1997.

....authorities. These subnetworks are called domains or Autonomous Systems. This way, the topology of the Internet can be studied at two di erent granularities. At the router level, we represent each router by a node [43] At the inter domain level, each domain is represented by a single node [22] and each edge is an inter domain interconnection. The study of the topology at both levels is equally important. The Internet community develops and employs di erent protocols inside a domain and between domains. An intra domain protocol is limited within a domain, while an inter domain protocol ....

....along the shortest path between the two nodes. Most studies report minimum, maximum, and average values and plot the degree and distance distribution. We denote the number of nodes of a graph by N , the number of edges by E, and the diameter of the graph by . Using these metrics Govindan and Reddy [22] study the growth of the inter domain topology of the Internet between 1994 and 1995. The graph is sparse with 75 of the nodes having degrees less or equal to two. Pansiot and Grad [43] study the topology of the Internet in 1995 at the router level. The distances they report are approximately two ....

R. Govindan and A. Reddy. An analysis of Internet Inter-domain topology and route stability. Proc. IEEE INFOCOM, Kobe, Japan, April 7-11 1997.

....knowledge about the Internet structure, the research on the subject has recently produced many contributions. More specifically, there is a wide research area focusing on the discovery of the topology underlying the Internet structure, either at the AS and at the router level (see, for example, [6], 7] 8] Other researchers concentrate more directly on the above mentioned relationships and on the hierarchy that they induce Work partially supported by European Commission Fet Open project COSIN COevolution and Self organisation In dynamical Networks IST2001 33555, by Progetto ....

.... Progetto ALINWEB: Algoritmica per Internet e per il Web , MIUR Programmi di Ricerca Scientifica di Rilevante Interesse Nazionale, and by The Multichannel Adaptive Information Systems (MAIS) Project , MIUR Fondo per gli Investimenti della Ricerca di Base. on the set of ASes. Govindan and Reddy [6] study the interplay between the degree of the ASes and their rank in the hierarchy, where the degree of an AS is the number of ASes that have some kind of relationship with it. Gao [1] studies, for the first time, the following problem. ASes are the vertices of a graph (AS graph) where two ASes ....

R. Govindan and A. Reddy, "An analysis of internet inter-domain topology and route stability," in Proc. IEEE INFOCOM 1997.

....It is this graph that we will use to study the router level Internet topology. Previous work. The neighbourhood of a node is important for estimating the complexity of various networking protocols such as DVMPR and QoSMIC [3, 16] There have been several measurements of the Internet topology [6, 10, 7]. These studies focus on the collection of data while the analysis appears secondary. There has not yet been a comprehensive study of the Internet topology at the router level. In contrast, the interdomain level has been studied lately [15] In a parallel tangent, several people have studied ....

R. Govindan and A. Reddy. An analysis of internet inter-domain topology and route stability. Proc, IEEE INFOCOM, Kobe, Japan, April 7-11 1997.

....and in part by an ARO MURI project, Mathematics of Failures in Complex Systems . of network components, service provider equipments, and user accessing facilities. Govindan et al. revealed that both the route availability and the mean reachability duration have degraded with the Internet growth [1]. Long et al. evaluated mean time to failure (MTTF) mean time to repair (MTTR) and availability and reliability of a sample of hosts by repeatedly polling the hosts and discovered that daily and weekly shutdowns appeared very commonly in the Internet [2] By periodically collecting data on a set ....

R. Govindan and A. Reddy, "An analysis of Internet inter-domain topology and route stability," in INFOCOM'97, 1997, pp. 850--857.

....such as OSPF, by reducing the control traffic both within areas as well as across the backbone. To verify this expectation, we compared ALVA with OSPF in several simulations. Simulation were performed using random graphs with 100 nodes. Nodes had an average degree of approximately 3. Recent work [GR97] shows that this is a realistic node degree for internetworks. The topologies were produced according to two general schemes. According to the first scheme, there is a backbone with 56 nodes, one area with 30 nodes, and 14 stub areas with one node each. We chose to use stub areas because we were ....

R. Govindan and A. Reddy. An analysis of internet inter-domain topology and route stability. In Proc. IEEE INFOCOM '97, Kobe, Japan, April 7-11 1997.

....They also find that a greedy algorithm for mirror placement (one which simply iteratively chooses the best node as the site for the next mirror) performs better than a tree based algorithm. Both router level and inter domain topology have been studied over the past five years [21] 22] 23] [24], 25] Our clustering algorithm uses BGP data thus inter domain topology is most relevant to this work. In [24] Govindan and Reddy characterize inter domain topology and route stability using BGP routing table information collected over a one year period. In that work the authors describe ....

....as the site for the next mirror) performs better than a tree based algorithm. Both router level and inter domain topology have been studied over the past five years [21] 22] 23] 24] 25] Our clustering algorithm uses BGP data thus inter domain topology is most relevant to this work. In [24], Govindan and Reddy characterize inter domain topology and route stability using BGP routing table information collected over a one year period. In that work the authors describe inter domain topology in terms of diameter, degree distribution and connectivity characteristics. Inter domain routing ....

R. Govindan and A. Reddy, "An analysis of internet inter-domain topology and route stability," in Proceedings of IEEE INFOCOM '97, Kobe, Japan, April 1997.

....and a continuous expansion of the underlying Internet connectivity. Researchers have been interested in how such individually and locally made engineering efforts to optimize network performance are reflected in the current global Internet connectivity. A number of recent empirical studies [9, 5, 15, 17] have tried to characterize the spatial properties and the temporal development of the Internet topology. For the purpose of this paper, Internet topology is defined on two different abstract levels: router level and Autonomous System (AS) level. An AS is a network under a single administration ....

R. Govindan and A. Reddy. An analysis of Internet inter-domain topology and route stability. In Proceedings of IEEE Infocom, 1997.

....and an edge represents the adjacency of two routers. Approaches to generating AS graphs can be grouped into three categories: AS Path based graph, traceroute based graph, and synthetic graph. A. AS Path based AS graph The AS graph is derived from the AS Path information in BGP Update messages [6] or in BGP tables [14] Each element of an AS Path defines a node and each successive pair of domains in the AS Path represents the endpoints of an edge. Although simple, this approach may not yield a complete AS graph due to the unavailability of the complete BGP routing information. Also, the AS ....

Ramesh Govindan and Anoop Reddy. An analysis of Internet inter-domain topology and route stability. In Proc. 16th IEEE INFOCOM, April 1997.

....These contractual commercial agreements between administrative domains play a crucial role in shaping the structure of the Internet and the end to end performance characteristics. Previous work on the Internet topology has been focused on the interconnection structure at either AS or router level [4, 5, 6, 7, 8, 9]. Since routing between ASes is controlled by BGP a policy based routing protocol, connectivity does not imply reachability. For example, national ISPs A and B are connected to their customer a regional ISP C respectively. Although ISPs A and B are connected through ISP C, ISP A can not reach ....

....implied by AS relationships and derive routing table entry patterns as the result of routing policies. We then infer the AS relationships based on the heuristic that the size of an AS is typically proportional to its degree in the AS graph. This heuristic has been used by Govindan and Reddy [5] in classifying ASes into four levels of hierarchy. Our heuristic algorithms classify an interconnected AS pair into having a provider customer, peering, or sibling relationship. The running time of the algorithm is linear in the total number of consecutive AS pairs in the routing tables. To the ....

[Article contains additional citation context not shown here]

R. Govindan and A. Reddy, \An analysis of Internet inter-domain topology and route stability," in Proc. IEEE INFOCOM, April 1997.

....that improve the similarity of generated graphs with real AS graphs, and first results are encouraging. Note that our structural explanation of P , Q, R and I nodes is roughly related to the hierarchical classification of nodes in the transit stub model [3] to the degree based classification in [12], and to the classification into core, subcore, regional, and stub nodes in [8] In [12] AS nodes are classified into national or international backbones (degree 28) large regional providers (10 27) large MAN providers (4 9) and multi campus corporate or academic networks (1 3) This ....

....are encouraging. Note that our structural explanation of P , Q, R and I nodes is roughly related to the hierarchical classification of nodes in the transit stub model [3] to the degree based classification in [12] and to the classification into core, subcore, regional, and stub nodes in [8] In [12], AS nodes are classified into national or international backbones (degree 28) large regional providers (10 27) large MAN providers (4 9) and multi campus corporate or academic networks (1 3) This classification is becoming outdated by the commercialization of the Internet over the past few ....

R. Govindan and A. Reddy, "An analysis of internet inter-domain topology and route stability," in Infocom'97, 1997.

....10000 nodes) is impossible today. While it is easy to obtain a list of all the ASs, their connectivity depends on local public and private peering arrangements and the routing policies put in place by ISPs. Some heuristic methods to infer subgraphs of the full AS topology are described in [16, 8, 13]. 2.2 State Next, we compare inferring the state of the network, assuming that the underlying topology is known. Network state includes the operational state of links and routers, the routing and forwarding tables in effect, and other variables that do not directly depend on traffic (e.g. ....

Ramesh Govindan and Anoop Reddy. An analysis of Internet inter-domain topology and route stability. In Proc. IEEE INFOCOM '97, April 1997.

....of F with respect to the proactive and reactive performance measures entails studying its dependence on topology G, the size of the filter net T , its structure, and routing R. 4. 2 Power law Network Topology Empirical evidence shows that Internet AS topology exhibits power law connectivity [6, 10] which may also hold for router topologies [18] Power law graph structure induces centers where connectivity is concentrated on a few nodes, with most vertices possessing sparse connectivity (e.g. comprised of AS stubs and non transit multi homed AS s) A key aspect of our DDoS benchmark ....

R. Govindan and A. Reddy. An analysis of Internet inter-domain topology and route stability. In Proc. IEEE INFOCOM '97, 1997.

....the router graph represents a router [13] 1] 2] 5] 15] and an edge represents the adjacency of two routers. Approaches to generating AS graphs can be grouped into three categories: 1. AS Path based AS graph. The AS graph 2 is derived from the AS Path information in BGP Update messages [6] or in BGP tables [14] Each element of an AS Path defines a node and each successive pair of domains in the AS Path represents the endpoints of an edge. Although simple, this approach may not yield a complete AS graph due to the unavailability of the complete BGP routing information. Also, the AS ....

Ramesh Govindan and Anoop Reddy. An analysis of Internet inter-domain topology and route stability. In Proc. 16th IEEE INFOCOM, April 1997.

....It is this graph that we will use to study the router level Internet topology. Previous work. The neighbourhood of a node is important for estimating the complexity of various networking protocols such as DVMPR and QoSMIC [3, 16] There have been several measurements of the Internet topology [6, 10, 7]. These studies focus on the collection of data while the analysis appears secondary. There has not yet been a comprehensive study of the Internet topology at the router level. In contrast, the interdomain level has been studied lately [15] In a parallel tangent, several people have studied ....

R. Govindan and A. Reddy. An analysis of internet inter-domain topology and route stability. Proc. IEEE INFOCOM, Kobe, Japan, April 7-11 1997.

....paper is to analyze and quantify the e ect of congestion on the stability of two common Internet routing protocols OSPF and BGP when the routing protocol messages are not isolated from data trac. A number of previous studies have dealt with various aspects of routing behavior in the Internet [3, 7, 12, 13, 17]. These studies used experimentation to measure and document various aspects of the routing dynamics, and they identify several stability related problems in today s Internet. Some of these studies have also tried to identify the root causes of these problems, attempted to analyze them, and ....

....routing uctuations in the NSFNET originated at the edge of the network, and had cycle intervals of a few minutes. Paxson [17] used the traceroute utility at 37 Internet sites to analyze the routing behavior for pathological conditions, routing stability and routing symmetry. Govindan and Reddy [7] used a year s worth of inter domain routing traces collected in 1994 95 and analyzed the Internet inter domain topology, its routing stability behavior, and the e ect of growth on these two characteristics. One of the key ndings of their study was that the stability behavior of Internet routes ....

R. Govindan and A. Reddy. An Analysis of Internet Inter{Domain Topology and Route Stability. In Proceedings of the IEEE INFOCOM '97, 1997.

....size of the lter net T , its structure, and routing R. Our goal is to uncover the above relationships and nd feasible means to economize T when trying to achieve a target performance. 4. 2 In uence of Topology Empirical evidence shows that Internet AS topology exhibits power law connectivity [6, 10] which may also hold for router topologies [18] Power law graph structure induces centers where connectivity is concentrated on a few nodes, with most vertices possessing sparse connectivity (e.g. comprised of AS stubs and non transit multi homed AS) A key aspect of our DDoS benchmark ....

R. Govindan and A. Reddy. An analysis of Internet inter-domain topology and route stability. In Proc. IEEE INFOCOM '97, 1997.

....relationships with other ISPs. The analysis of these business relationships reveals important aspects of their business models. 1 Introduction The analysis of the Internet is not sufficient anymore if it is solely based on the link state between routers (i.e. Internet topology analysis) [1][4] There are two main reasons. First, the Internet is becoming an integrated services network. More and more services are being introduced that require a better than best effort network service quality. That means, in addition to the connectivity analysis, the performance of routes through the ....

Govindan, R. and A. Reddy, "An Analysis of Internet Inter-Domain Topology and Route Stability", Proceedings of the IEEE Infocom, Japan (1997).

....result of the fact that the probability of a search success increases exponentially with the number of domains that have members. It also results from the fact that the diameter of the Internet is only 10 domain hops. It has been 10 domain hops for the past 6 years and is unlikely to change [3,9]. Therefore, even for the case of unpopular groups when we send the packets to the home domain, on average the home domain is only 5 domain hops away. Although the 2 domain hops curve in Figure 12 shows a worst case inefficiency of 1.23, the occurrence of the worst case behavior is unlikely in ....

R. Govindan and A. Reddy, "An Analysis of Internet Inter-Domain Topology and Route Stability," Technical report USC-CS-96-642, Department of Computer Science, University of Southern California, Proc. IEEE INFOCOM'97 (1997).

....of measurement based studies of BGP protocol traffic and theoretical analysis of BGP convergence properties. Extensive traces of BGP update messages have been used to characterize the structure (and growth) of the Internet topology, as well as the stability of routes to destination prefixes [10 14]. In contrast, research on BGP convergence has focused on determining what combination of BGP policies would cause a group of ASes to continually advertise and withdraw routes to a given prefix [3, 4, 15 17] BGP convergence problems would not arise if every AS selects shortest path routes. ....

R. Govindan and A. Reddy, "An analysis of Internet inter-domain topology and route stability," in Proc. IEEE INFOCOM, April 1997.

....These contractual commercial agreements between administrative domains play a crucial role in shaping the structure of the Internet and the end to end performance characteristics. Previous work on the Internet topology has been focused on the interconnection structure at either AS or router level [4, 5, 6, 7, 8, 9]. Since routing between ASes is controlled by BGP a policy based routing protocol, connectivity does not imply reachability. For example, national ISPs A and B are connected to their customer a regional ISP C respectively. Although ISPs A and B are connected through ISP C, ISP A can not ....

....implied by AS relationships and derive routing table entry patterns as the result of routing policies. We then infer the AS relationships based on the heuristic that the size of an AS is typically proportional to its degree in the AS graph. This heuristic has been used by Govindan and Reddy [5] in classifying ASes into four levels of hierarchy. Our heuristic algorithms classify an interconnected AS pair into having a provider to customer, peer to peer, or sibling to sibling relationship. The running time of the algorithm is linear in the total number of consecutive AS pairs in the ....

[Article contains additional citation context not shown here]

R. Govindan and A. Reddy, "An analysis of Internet inter-domain topology and route stability," in Proc. IEEE INFOCOM, April 1997.

....a multicast receiver site is located only at each intersection of the grid in a network. Thus, the network model is a non hierarchical, flat structure. It is predicted that the future Internet will tend towards a non hierarchical, flat network as the number of domains increases (see [9] for more details) The number of nodes in a network, 2 N , represents the network size where N is the number of nodes in each edge of a square network. To assign links between nodes, the link assignment model proposed by Waxman (see [14] is applied in which the probability of there being ....

R. Govindan and A. Reddy, "An Analysis of Internet Inter-Domain Topology and Route Stability," Proceedings of IEEE INFOCOM, pp. 850 - 857, 1997.

....algorithm for the QoS partition problem with general discrete cost functions. Going back to the example of figure 1, n = 3; m = 3 3 2 = 8, and for ffl = 1 mn 3 log m ffl = 64800. In general, the number of DS domains that a connection traverses in the US is almost always below 5 [GR97] Assuming a full utilization of the DS field in the IP header enables to support 256 service classes, which translates to m = 256 5 = 1280, we get for ffl = 1 mn 3 log m ffl = 1:6E8. Although this is well within the computing capabilities of todays hardware, improving the algorithm ....

Ramesh Govindan and Anoop Reddy. An analysis of Internet inter-domain topology and route stability. In IEEE INFOCOM'97, pages 851 -- 858, April 1997.

....of measurement based studies of BGP protocol traffic and theoretical analysis of BGP convergence properties. Extensive traces of BGP update messages have been used to characterize the structure (and growth) of the Internet topology, as well as the stability of routes to destination prefixes [7 10]. In contrast, research on BGP convergence has focused on determining what combination of BGP policies would cause a group of ASes to continually advertise and withdraw routes to a given prefix [3, 4, 11, 12] BGP convergence problems would not arise if every AS selects shortest path routes. ....

R. Govindan and A. Reddy, "An analysis of Internet inter-domain topology and route stability," in Proc. IEEE INFOCOM, April 1997.

....such as OSPF, by reducing the control traffic both within areas as well as across the backbone. To verify this expectation, we compared ALVA with OSPF in several simulations. Simulations were performed using random graphs with 100 nodes. Nodes had an average degree of approximately 3. Recent work [16] shows that this is a realistic node degree for internetworks. The topologies were produced according to two general schemes. According to the first scheme, there is a backbone with 56 nodes, one area with 30 nodes, and 14 stub areas with one node each. We chose to use stub areas because we were ....

R. Govindan and A. Reddy, "An analysis of internet inter-domain topology and route stability," in Proc. IEEE INFOCOM '97, Kobe, Japan, April 7-11 1997.

....in Figure 1. These subnetworks are called domains or autonomous systems 2 . This way, the topology of the Internet can be studied at two different granularities. At the router level, we represent each router by a node [14] At the inter domain level, each domain is represented by a single node [10] and each edge is an inter domain interconnection. The study of the topology at both levels is equally important. The Internet community develops and employs different protocols inside a domain and between domains. An intra domain protocol is limited within a domain, while an inter domain protocol ....

....shortest path between the two nodes. Most studies report minimum, maximum, and average values and plot the outdegree and distance distribution. We denote the number of nodes of a graph by N , the number of edges by E, and the diameter of the graph by ffi. Real network studies. Govindan and Reddy [10] study the growth of the inter domain topology of the Internet between 1994 and 1995. The graph is sparse with 75 of the nodes having outdegrees less or equal to two. They distinguish four groups of nodes according to their outdegree. The authors observe an increase in the connectivity over time. ....

[Article contains additional citation context not shown here]

R. Govindan and A. Reddy. An analysis of internet interdomain topology and route stability. Proc. IEEE INFOCOM, Kobe, Japan, April 7-11 1997.

....the session will use the same route at a later instant in time. However, he notes further that, there is extensive variability in the measure from site to site (variability of 50 60 ) and therefore, while one end to end path may remain fairly stable, another might be less so. Govindan et al.[4] arrive at a similar conclusion, based on their analysis of inter domain routing protocol traces gathered at various points in the Internet backbone over 18 month period. More recently, Labovitz et al. 8] through measurements of inter domain route updates at various network access points, ....

R. Govindan and A. Reddy. An analysis of internet inter-domain topology and route stability. In Proceedings of the IEEE INFOCOM, April 1997.

....the session will use the same route at a later instant in time. However, he notes further that, there is extensive variability in the measure from site to site (variability of 50 60 ) and therefore, while one end to end path may remain fairly stable, another might be less so. Govindan et al.[5] arrive at a similar conclusion, based on their analysis of inter domain routing protocol traces gathered at various points in the Internet backbone over 18 month period. More recently, Labovitz et al. 9] through measurements of inter domain route updates at various network access points, ....

R. Govindan and A. Reddy. An analysis of internet inter-domain topology and route stability. In IEEE Proceedings of the INFOCOM, April 1997.

....the session will use the same route at a later instant in time. However, he notes further that, there is extensive variability in the measure from site to site (variability of 50 60 ) and therefore, while one end to end path may remain fairly stable, another might be less so. Govindan et al.[3] arrive at a similar conclusion, based on their analysis of inter domain routing protocol traces gathered at various points in the Internet backbone over an 18 month period. More recently, Labovitz et al. 7] through measurements of inter domain route updates at various network access points, ....

R. Govindan and A. Reddy. An analysis of internet inter-domain topology and route stability. In IEEE Proceedings of the INFOCOM, Apr. 1997.

.... packet delays was conducted as early as 1971 [K76] More measurement studies [CPB93, CPB93 2, H90, PF95] were performed on the NSFNET after it replaced ARPANET in 1990 [MERIT] More recently measurement studies showing the Internet routing instability and dynamics have also been reported in [GR97, LMF97, P97]. These measurement studies not only exposed the unexpected behavior of the current Internet protocols, but also help us better understand the dynamics of large scale systems. End to end behaviors in the Internet, including delay and hop count (number of hops along a path from one host to ....

....would be the average hop count because of some good properties of hop count distribution. As shown in the previous section and in [CC96] hop count is pretty much evenly distributed with the mean at the center. So average hop count should give us a fairly good impression on how deep vertically [GR97] the Internet is. However, the Internet is so highly heterogeneous, and our measurement shows that routes to different countries are really country specific, one has to be careful at choosing sample hosts when doing such measurement. Another interesting problem is how the diameter or average ....

R. Govindan, and A. Reddy. An analysis of Internet inter-domain topology and route stability. In the Proceedings of the IEEE Infocom 1997.

No context found.

R. Govindan and A. Reddy. An Analysis of Internet Inter-Domain Topology and Route Stability. In Proc. IEEE INFOCOM '97, Kobe, Japan, Apr 1997.

....These traces are incomplete in parts, caused by failures at the trace collection points. In our analysis, we have been careful to eliminate the effect of these trace flaws. A macroscopic examination of the trace data reveals nearly linear growth in the number of prefixes, links, and domains [10]. On average, a new prefix appears in the routing system every 25 minutes, a new domain every 12 hours, and a new inter domain link every 8 hours. Of interest to us, however, are more detailed measures of growth in topology and changes in route stability. For this, we focused on three different ....

R. Govindan and A. Reddy. An Analysis of Internet InterDomain Topology and Route Stability. Technical Report 96-642, Computer Science Dept., University of Southern California, Nov 1996.

....traceroute targets. Techniques for collecting other representations of Internet, such as the AS (Autonomous System) topology, have also been documented in the literature. In one approach, traces of backbone routing activity over a period of several days have been used to infer inter AS links [12]. Each link represents an inter ISP peering or a customer ISP connection. Instantaneous dumps of backbone routing tables have also been used to infer AS level links [3] In some cases, router support can be used to determine router adjacencies. For example, Intermapper [7] builds a list of router ....

....exits interface B. Mercator therefore repeatedly sends alias probes to every known interface. To limit probing traffic, only a small number of alias probes are run concurrently. The efficacy of such repeated alias probing is based on the observation that there exist dominant paths [14] and routes [12] in the Internet. There is a high likelihood that, eventually, a large fraction of the alias probes will be returned via the router s dominant path to the sending host. The second refinement uses source routed alias probes. Figure 3(b) explains why this is necessary. In practice, large backbones ....

R. Govindan and A. Reddy. An Analysis of Internet Inter-Domain Topology and Route Stability. In Proc. IEEE INFOCOM '97, Kobe, Japan, Apr 1997.

....and inter AS connectivity was determined by bi lateral commercial agreements on tra c exchange. Corresponding to the explosive Internet growth, the number of ASes has grown rapidly since then. Between mid 1994 and late 1995, the number of ASes grew at an astonishing rate of nearly two a day [11]. At the time of writing, there were more than 3000 1 ASes in the Internet. Despite the reduced constraints on AS interconnections, reference [11] shows, from a study of routing table traces, that ASes can be classi ed by their degree in the inter AS topology into four levels. At the top level ....

....the number of ASes has grown rapidly since then. Between mid 1994 and late 1995, the number of ASes grew at an astonishing rate of nearly two a day [11] At the time of writing, there were more than 3000 1 ASes in the Internet. Despite the reduced constraints on AS interconnections, reference [11] shows, from a study of routing table traces, that ASes can be classi ed by their degree in the inter AS topology into four levels. At the top level are the national or international backbones; examples of these include backbones operated by MCI 2 , Advanced Networks and Services, and Sprint. At ....

R. Govindan and A. Reddy. An Analysis of Internet Inter-Domain Topology and Route Stability. In Proc. IEEE INFOCOM '97, Kobe, Japan, Apr 1997.

No context found.

Ramesh Govindan and Anoop Reddy. An analysis of internet inter-domain topology and route stability. In INFOCOM (2), pages 850--857, 1997. 41

No context found.

R. Govindan and A. Reddy. An Analysis of Internet Inter-Domain Topology and Route Stability. In IEEE InfoCom, 1997. 18

No context found.

R. Govindan and A. Reddy. An analysis of internet interdomain topology and route stability. In INFOCOM, pages 850--857, 1997.

No context found.

R. Govindan and A. Reddy, "An analysis of internet inter-domain topology and route stability," in Proc. IEEE INFOCOM'97, Apr. 1997, pp. 851--858.

No context found.

Ramesh Govindan and Anoop Reddy. An Analysis of Internet Inter-Domain Topology and Route Stability. In Proceedings of INFOCOM 1997.

No context found.

R. Govindan and A. Reddy. An Analysis of Internet Inter-Domain Topology and Route Stability. In IEEE InfoCom, 1997. 18

No context found.

R. Govindan and A. Reddy, "An analysis of internet inter-domain topology and route stability," in Proceedings of IEEE INFOCOM '97, Kobe, Japan, April 1997.

No context found.

R. Govindan and A. Reddy, \An analysis of Internet inter-domain topology and route stability," in Proc. IEEE INFOCOM, April 1997.

No context found.

R. Govindan and A. Reddy, "An analysis of Internet inter-domain topology and route stability," in INFOCOM'97, 1997, pp. 850--857.

No context found.

R. Govindan and A. Reddy, "An analysis of internet inter-domain topology and route stability," in Proceedings of INFOCOM'97, April 1997.

No context found.

Govindan R., Reddy A.: An analysis of Internet Inter-Domain Topology and Route, Stability. In Proc. IEEE INFOCOM '97, Kobe, Japan, Apr 1997.

No context found.

R. Govindan and A. Reddy, "An Analysis of Internet Inter-Domain Topology and Route Stability," Technical report USC-CS-96-642, department of computer science, University of Southern California, In Proc of INFOCOM'97, 1997.

No context found.

R. Govindan and A. Reddy. An analysis of Internet inter-domain topology and route stability. In Proc. IEEE INFOCOM, April 1997.

First 50 documents  Next 50

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC