Past Projects



A Framework for Manageability in Future Routing Systems. This was a joint project with the University of Minnesota (Prof. Zhi-Li Zhang) and the University of Massachusetts (Prof. Lixin Gao) funded by NSF under the FIND initiative (NSF grant CNS-0627004). The project addressed fundamental questions on building manageability into routing systems for future Internet architectures. Its goals were two-fold: i) develop a framework for specifying, understanding, and evaluating what features should/could be "designed-in" into routing systems in support of manageability; and ii) evaluate design choices and trade-offs thereof in terms of performance and manageability.

Project Website

Related Publications


A Market Approach to Controlling the Proliferation of Internet Routes. This was a project aimed at developing mechanism to address the growth of core Internet routing tables.

The growth of core Internet routing tables is obviously a direct consequence of the growth of the Internet itself. However, it has been proceeding at a pace that far exceeds that of the Internet growth, and this is threatening the scalability of the Internet routing system, both in the short-term (most routers have hard limits on the number of routes they can store) and in the long-term (it out-paces the rate of improvement from technology itself). The main culprit behind this trend is the desire for individual “sites” to improve the reliability of their Internet connectivity; clearly a natural desire given the growing dependency on Internet connectivity. This desire for better reliability commonly manifests itself through multi-homing decisions, i.e., connectivity to the Internet through more than one provider to enjoy the benefits of greater path diversity, and herein lies the problem. When sites are multi-homed, their individual address sub-block are advertised by the different providers to which they connect, and this typically affects the ability of upstream providers to aggregate those routes; hence creating additional entries in core Internet routing tables.

In this project, we investigated market mechanisms to better redistribute the cost of additional routing entries caused by multi-homing to the sites responsible for those new entries. The hope is that this redistribution can help control the growth of routing tables or at least ensure that their growth is borne by those that are causing it and benefit from it.

The project also seeks to more broadly explore the use of market mechanisms to tackle a number of technical networking problems, including aspects related to promoting the adoption of those solutions.

Related Publications:



Distributed Uplink Scheduling in CDMA Networks. This was a project in collaboration with Ashwin Sridharan and partially funded by Sprint Labs, which explored issues that arise in CDMA networks when devices (mobiles) are afforded some level of independence in making transmission decisions, instead of being under the tight control of a base station. Giving mobile devices some flexibility in deciding when to transmit and at what rate is increasingly desirable because of the diversity of applications they are now capable of running and that exhibit a broad range of communication requirements. However, allowing devices to make individually controlled transmission decisions may affect global system performance, and one of the goals of this project is to explore this trade-off in a number of different settings.

Related Publications


Towards Large-Scale Flat Networks. This was a project sponsored by and in collaboration with Siemens. Flat networks like Ethernet have many advantages in terms of simplicity and flexibility, e.g., plug-&-play and little or no configuration requirements. However, they suffer from a number of potential scalability limitations, which have limited the scope of their deployments and promoted the use of (hierarchical) routed solutions to build large networks, the Internet being a case in point. This work is not aimed at replacing the Internet with one big flat network, but it explores various issues aimed at improving the scalability of flat networks. In particular, it targets two important factors that affect scalability: (i) reliance on broadcast for address discovery; and (ii) loop prevention during path changes.

Related Publications


Data plane aggregation. This work was supported through NSF grant ITR-0085930 and aimed at developing a better understanding of the relations that exist between QoS provided as some aggregate level, e.g., a service class as in the Differentiated Services model, and the actual QoS that individual users experience. Of particular interest are models that allow explicit evaluation of individual QoS measures, and their use in identifying characteristics of user traffic that can result in significant differences between individual and aggregate QoS measures. This led to the development of models that allow the evaluation of the loss probability experienced by individual connections and when and why it differs from the aggregate loss probability. The environment that is assumed consists of a single FIFO queue where all the individual users belonging to the same service class are multiplexed. A separate but related perspective is that of security, namely, understanding the extent to which a single (or a few) user can affect the performance of many other users. In that context, we investigated the extent to which more sophisticated attack schemes can defeat existing mechanisms, and used that understanding towards developing better defenses.

Related Publications


Traffic engineering in IP networks. This work was aimed at better understanding what is achievable using different traffic engineering techniques, and more important how one can devise solutions that are not highly sensitive to the quality of the “input” on which traffic engineering decisions are based. One problem that was investigated included how to best group traffic flows to minimize the number of distinct paths that had to be established in order to achieve optimal performance, and how this grouping affected both long term and short term performance. Another problem focused on devising minor modification to existing IP forwarding in order to allow the implementation of near optimal traffic distribution over existing IP networks. Yet another area of investigation involved evaluating traffic engineering solutions that can provide resilience to changes caused by link or node failures or by fluctuations in traffic patterns. This work was supported in part by NSF grants ANI-9902943 and ITR-0085930, and by a gift from Sprint ATL, and was partly carried out in collaboration with Christophe Diot (now at Thomson Paris Research Lab) and his former group at Sprint ATL. Another aspect of this work, supported by supported by NSF grant ITR-0085930, investigated the use of overlay networks to deliver better service “guarantees” over IP networks by leveraging the (path) diversity offered by the availability of a large number of peers. Part of this investigation was carried in the context of real-time applications such as VoIP and video (see above), but another area of focus was to develop simple techniques for quickly identifying peers that could act as “good” relay nodes for alternate overlay paths, i.e., overlay paths with performance that was largely uncorrelated with that of the default path. The challenge was to retain the benefits of having access to a very large number of possible choices, while devising simple and scalable solutions, i.e., that required little processing and only minimal storage.

Related Publications


Robust and flexible scheduling mechanisms. The work explored the development of a number schedulers aimed at increasing the flexibility with which bandwidth can be distributed across users with different types of reservations. One such direction involved supporting bandwidth guarantees with preemption capability. Another direction involved provision of delay guarantees while allowing for the transmission of excess traffic. A software implementation of the ADQ scheme that supports this capability is available for download from the “Softwarepage of the Multimedia and Networking Lab. Some of this work has been supported in part by NSF grant ANI-9902943 and by a grant from Nortel Networks.

Related Publications


Building flexible ad-hoc and wireless networks. This work looked at a number of issues that arise in the context of wireless ad hoc networks, and in particular the trade-off that exists between the flexibility that wireless links afford and the complexity they introduce, in part because of this very same flexibility that creates both opportunities for adaptation and alternative configurations and challenges associated with controlling these capabilities. One of the first problems we investigated was aimed at gaining a better understanding of whether or not the Bluetooth technology represents a viable option for building ad-hoc networks. In particular, the master-slave design of the Bluetooth protocol embodies a specific choice between enabling “spontaneous” communications between neighboring devices and providing control on how such communications are to take place. This design choice has implications for the complexity of building large-scale ad hoc networks with Bluetooth, because of the challenges it introduces in the topology construction phase that is required to enable end-to-end connectivity between nodes. Another topic we investigated is how to exploit the many distinct transmission resources, e.g., different channels, that wireless technology offers, as a means for improving overall transmission performance. In particular, we have attempted to develop an understanding of when, why, and how channel diversity can yield significant improvements in achievable transmission rates. In this context, our focus has been on exploiting the use of diversity at the “network” level rather than using more traditional “physical” layer approaches.   This work was supported in part by NSF grants ANI-9902943 and ITR-0085930.

Related Publications


Evaluation of the robustness (and usefulness) of some QoS services. This work was based on both analysis (mostly simulations in this case), and experimentation on a testbed that consisted of routers with various Quality-of-Service (QoS) capabilities. In addition, a connection to the QBone (a QoS enabled portion of the Internet2 network) was used to complement the local testbed measurements with a wide area experiments. Of particular interest was the impact that network induced perturbations can have on conformance checks performed at boundaries between provider domains.  This work was supported in part by a grant from Nortel Networks and by NSF grant ANI-9906855.

Related Publications


Application versus network level performance. This work aimed at a better understanding of how changes in network performance (and service parameters) affect the performance seen by applications and was initially supported by NSF grant ANI-9906855. Subsequent extensions have been supported through NSF grant ITR-0085930. One suite of experiments involved passing packet video streams through policers with different combination of parameters, and evaluating (quantitatively and qualitatively) the evolution of application level performance. The quantitative evaluation of video quality was done using the VQM tool developed by The Institute for Telecommunications Science and was carried out over both a local testbed and the wide area QBone testbed. Testing over the QBone infrastructure was done in collaboration with members of the PennNet and Computing departments of the University of Pennsylvania (see also their Internet2 related activities) and with researchers at iCAIR and the IBM T.J. Watson Research Center. Additional experiments were conducted for both voice and video application in collaboration with other researchers at UMass and UMinn. over a wide-area testbed connecting all three sites. Exploring how to accurately and easily monitor the quality of video transmissions over packet networks was also investigated in collaboration with John Apostolopoulos from HP Labs.


Another set of experiments carried out with collaborators at
AT&T Research, Florham Park, NJ, focused on the use of aggregate and "non-intrusive" performance measures for estimating the actual throughput experienced by TCP based applications. Non-intrusive refers to measures that are readily available from routine network monitoring, e.g., from routers MIBs, and do not require any flow specific awareness. The work involved the development of models (modification of existing models) that were capable of accurately predicting TCP throughput on the basis of such information, and the evaluation of their accuracy using both simulations and testbed experiments.

Related Publications


Control plane aggregation. This work looked at when and where it is useful to aggregate reservations in order to minimize the amount of state and processing that needs to be performed in backbone routers. The focus was on developing an understanding of the storage and message cost of different reservation aggregation rules in the context of Internet-like topologies. The work involved both the development of efficient aggregation algorithms, and their evaluation, primarily by means of simulations, in realistic settings. The work has been supported in part through NSF grant ITR-0085930.

Related Publications