Active networks is a relatively new concept. In an active network, network nodes in addition to forwarding packets, perform customized computation on the messages flowing through them. In this thesis, we study resource allocation in the context active networks. Resources in an active network node include both CPU and bandwidth.

A key feature of active networks is the capability to dynamically deploy services. In this paper, we present a scheme to classify service deployment mechanisms of existing or future active network architectures. Distributed algorithms (services), as being implemented in active networks, can be described based on active packets or as distributed programs running on active nodes. Although both programming models are basically equivalent, some services are more naturally implemented in either way. This paper proposes an active node architecture that supports the implementation and deployment of services according to both programming models. We point out that a combination of in-band and out-of-band service deployment is needed to dynamically deploy services implemented in either model. Furthermore, we argue that composing services from service logic implemented in either programming model is beneficial for the design of efficient and flexible services. We reason that a service abstraction in the form of a service description language is necessary to cope with real world scenarios.

In this paper, we examine the application of active networking technology to future mobile networks. We first introduce an architecture for programmable 4th generation (4G) mobile networking, including all system layers on the network and terminal side. Based on this architecture, we discuss programmability in future mobile networks. We investigate the main driving forces and obstacles for the application of active networks. In particular, we show that flexible component installation and cross-layer interfaces are a main motivation for programmable mobile networks. This is illustrated by a number of applications for future mobile networks, including context-aware mobility management and paging, where flexibility is a key requirement for future mobile services.

Next-generation network architectures will be governed by the need for flexibility. Heterogeneous end-systems, novel communication abstractions, and security and manageability challenges will require networks to provide a broad range of services that go beyond the simple store-and-forward capabilities of today’s Internet. The proposed work introduces new abstractions for information transfer and data services in the network, which overcome the constraints of the end-to-end argument that has dominated current network design. The explicit separation of communication and processing allows the composition of a variety of information transfer patterns that expand the capabilities of the network to meet its next-generation challenges. Our key idea is to base communication abstractions on the transfer of information rather than the process of sending data, because the fundamental task of a network is to transfer information form one system to another. The process of handling data representation, transfer, and storage of information is considered a service. The proposed research will explore these new end-to-end information transfer and data service (ITDS) abstractions and pursue four goals: 1. Development and Prototyping of ITDS Abstractions: We propose to develop a prototype architecture that implements the main ideas of our ITDS abstractions. We propose

In this paper, we present a pattern based approach to network level service deployment for active networks. Network level service deployment includes discovering and selecting active nodes that fulfill the service requirements and are capable of running service components. Once a suitable set of active nodes is identified, service installation is triggered on those nodes. Our approach provides a first step towards composing service-specific deployment protocols from reusable protocol building blocks. We make extensive use of the distributed processing capability available in active networks. Our work is particularly well suited for networks where topology and available resources change frequently. Moreover, our approach is tailored to networks that potentially support a broad range of service types, as is the case for active networks. In such networks, it is inefficient or even impossible to constantly distribute all potentially useful node status data. Therefore, we exclusively gather information relevant to the deployment of the service in question. First simulations show promising results.

Tools have been developed to automatically integrate and test networking systems in reconfigurable hardware. These tools dynamically generate circuits for Field Programmable Gate Arrays (FPGAs). A library of hardware-accelerated modules has been developed that processes Internet Protocol (IP) packets, performs header rule matching, scans packet payloads, and implements per-flow queueing. Other functions can be added to the library as extensible modules. An integration tool was developed to enable a network administrator to specify how a customized system should examine, drop, bu#er, and/or modify packets. This tool joins together modules from the library to create a composite circuit that performs multiple functions. The tool allows additional modules to be quickly added to the library and integrated into systems. The integration tool has been used to create circuits that perform Internet firewall, network intrusion detection, network intrusion prevention, and Denial of Service (DoS) attack protection functions.

The flexibility to adapt to new services and protocols without changes in the underlying hardware is and will increasingly be a key requirement for advanced networks. Introducing a processing component into the data path of routers and implementing packet processing in software provides this ability. In such a programmable router, a powerful processing infrastructure is necessary to achieve a level of performance that is comparable to custom silicon-based routers and to demonstrate the feasibility of this approach. This work aims at the general design of such programmable routers and, specifically, at the design and performance analysis of the processing subsystem. The necessity of programmable routers is motivated, and a router design is proposed. Based on the design, a general performance model is developed and quantitatively evaluated using a new network processor benchmark. Operational challenges, like scheduling of packets to processing engines, are addressed, and novel algorithms are presented. The results of this work give qualitative and quantitative insights into this new domain that combines issues from networking, computer architecture, and system design.

In this paper, we present five case studies of advanced networking functions and how a network processor (NP) can provide high performance, and in particular the necessary flexibility compared with Application-Specific Integrated Circuits (ASICs). We first review the basic NP system architectures, and describe in more detail the IBM PowerNP architecture from a data plane as well as from a control plane point of view. We introduce models for the programmer's views of NPs that facilitate a global understanding of NP software programming. Then, for each case study, we present results from prototypes as well as general considerations that also apply to a wider range of system architectures. Namely, we investigate the suitability of NPs for quality-of-service (active queue management and tra#c engineering) , header processing (GPRS tunneling protocol) , intelligent forwarding (load-balancing without flow disruption), payload processing (code interpretation and just-in-time compilation in active networks), and protocol stack termination (SCTP). Finally, we summarize the key features required by each case study, and make concluding remarks regarding the future of NPs.

Active networking, where network nodes perform customized processing of packets, is a rapidly expanding field of research. This paper investigates the realization of service management on a Virtual Active Network (VAN), the key concept in our active network management framework [11]. A VAN can be seen as a thin (software) layer in an active networking environment, which creates a generic service abstraction, offered by the provider to the customer. The VAN concept transforms a multi-domain situation into a single domain for a customer. The customer can install, run and supervise network services into the VAN, without interaction with the provider. In this paper, we describe a service management toolkit, which facilitates designing and managing services taking advantage of active networking technology for code reuse, code distribution, flexible event filtering, and programmable information aggregation. Further, we show how the service management toolkit is realized on ANET, an active networking testbed that we are in the process of building.

We present the architecture, design, and implementation of a Smart Environment for Network Control, Monitoring and Management (SENCOMM). SENCOMM uses active network technology to comprise a Management Execution Environment (SMEE), which coexists with other execution environments (EEs). Management applications, called smart probes, run in the SMEE. A probe and its data are mobile executable code that are delivered to the active node within an Active Network Encapsulation Protocol (ANEP) datagram.