«Sixth Framework Programme Priority FP6-2004-IST-4 Situated and Autonomic Communications (SAC) Project Number: FP6-IST-27489 Deliverable D1.4/5/6v1 ...»
Autonomic Network Architecture
Sixth Framework Programme
Situated and Autonomic Communications (SAC)
Project Number: FP6-IST-27489
ANA Blueprint – First Version
Autonomic Network Architecture
ANA - Autonomic Network Architecture
ANA Blueprint – First Version
Christophe Jelger (UBasel) Editor Stefan Schmid (NEC) Christian Tschudin (UBasel) Christophe Jelger (UBasel) Ghazi Bouabene (UBasel) Guy Leduc (ULg) Lorenzo Peluso (FOKUS) Manolis Sifalakis (ULancs) Marcus Schoeller (ULancs) James Sterbenz (ULanc/University of Kansas) Authors Martin May (ETHZ) Matti Siekkinen (UOslo) Rudolf Roth (FOKUS) Stefan Schmid (NEC) Tanja Zseby (FOKUS) Thomas Plagemann (UOslo) Vera Goebel (UOslo) David Hutchison (ULancs) Reviewers Public Dissemination level 31st December 2006 Contractual delivery date 15th February 2007 Delivery Date 1.0 Version
This document constitutes the first version of the ANA Blueprint Architecture. It describes the initial reference model of the Autonomic Network Architecture (ANA) developed within the EU FP6 IST Project 27489.
The Blueprint includes a definition of the basic abstractions, core concepts and communication paradigms of ANA, defines
interfaces that an actual implementation of a network node must provide in order to support the ANA architectural concepts, and the information flow model for the relevant control and management information of an autonomic network.
ANA Blueprint Architecture, Basic Abstractions, Communication Paradigms, ANA Node, Information Flows Executive Summary The goal of the ANA project is to explore novel ways of designing and building networks beyond legacy Internet technology. The ultimate goal is to design and develop a novel network architecture that can demonstrate the feasibility and properties of autonomic networking . As specified in the description of work , it is the intension of the project to address the self-* features of autonomic networking such as self-configuration, self-optimization, self-monitoring, self-management, self-repair, and self-protection.
Because autonomic networking is very recent research area and there does not yet exist an autonomic network architecture, the ANA project is doing significant spadework on this wide and challenging research topic.
Designing a network architecture implies identifying and specifying the design principles of the system being developed. The architecture defines the atomic functions and entities that compose the network and specifies the interactions which occur between these various building blocks.
As the first draft of the ANA architectural blueprint specification, this document describes the initial reference model of the ANA autonomic network architecture. It includes a definition of the basic abstractions and concepts of the ANA architecture and the communication paradigms. The document also includes the abstract interfaces that an actual implementation of a network node must provide in order to support the ANA architectural concepts, and defines the information flow model for the relevant control and management information of an autonomic network.
Table of Contents 1 Introduction
1.1 Motivation and Objectives
1.2 Scope of Deliverable
1.3 Structure of the document
3 Basic Abstractions and Paradigms
3.1.1 Background and Motivation
3.1.2 Compartment Definition
3.1.3 Basic Compartment Functions
3.2 Node Compartment
3.2.1 Private views inside the node compartment
3.2.2 The node compartment and control channel
3.2.3 Example of Communication setup
3.3 Network Compartments
3.3.1 Compartments and Layers
3.3.2 Layering Compartments
3.4 Communication Paradigms
3.4.1 Intra-Compartment Communication
3.4.2 Inter-Compartment Communication
4 Autonomicity in ANA
4.1 Information Management
4.2 Network Monitoring
4.2.1 Key monitoring concepts
4.3 Network Management
5 ANA Node
5.1 Introduction and scope
5.2 Overview of the ANA Node functionalities
5.3 ANA Node: architecture and components
5.3.1 The information dispatch framework (IDF)
5.3.2 Key-Val Repository
5.3.3 The bootstrap procedure (BP)
5.3.4 The MINMEX Controller (MC)
5.3.5 The Playground
5.4 MINMEX API
5.4.1 Level 2 API
5.4.2 Level 1 API
5.4.3 Level 0 API
5.4.4 Level -1 API
5.4.5 Example Scenario
5.4.6 Code samples for levels 2, 1 and 0
6 Information Management
6.1 Scope of Information Management
6.2 Information Flow Requirements
6.2.1 Information Hook
6.2.2 Information Flow Characteristics
6.3 Information Flow Concepts
6.3.1 Information Hooks
6.3.2 Interactions between ANA entities
6.3.3 Communication Setup Example
6.3.4 Supported Information Sharing Methods
1 INTRODUCTIONThis document is the first version of the ANA Blueprint. It consists of deliverables D1.4v1 (First ANA Blueprint specifications), D1.5v1 (Autonomic functional blocks), and D1.6v1 (Information flow data items and interface specifications) which together specify the core operation of ANA. This first version of the Blueprint was produced during the first 9 months (M4-M12) of architectural work in 2006 and is the basis for both the development of a prototype and the deployment of a testbed in 2007. This upcoming prototyping phase will be used to test, validate, and refine the architectural concepts presented in this document, and will eventually lead to a second and more mature Blueprint document that will be released in the second-half of 2008 (M30).
This first version of the Blueprint describes the initial reference model of the ANA autonomic network architecture. This includes a definition of the basic abstractions and concepts of the ANA architecture and the communication paradigms. The Blueprint also includes the abstract interfaces that an actual implementation of a network node must provide in order to support the ANA architectural concepts. Finally, the Blueprint defines the information flow model for the relevant control and management information of an autonomic network (e.g., for configuration, monitoring, routing, optimization, resilience) and the necessary interfaces.
1.1 Motivation and Objectives This section recaps the main drivers behind the Autonomic Network Architecture (ANA) and its high-level objectives.
The overall objective is to develop a novel network architecture and to populate it with the functionality needed to demonstrate the feasibility of autonomic networking. In order to avoid the same failings of network architectures developed in the past, the guiding principles behind the architectural work in ANA are achieving maximum flexibility and providing support for functional scaling by design.
The former indicates that the project members do not envision a “one-size-fits-all” network architecture for the various different types of networking scenarios (e.g., Internet, Sensor Networks, Mobile Ad hoc Networks, Peer-to-peer Networks). Instead, the aim of ANA is to provide an architectural framework that enables the co-existence and interworking between different network architectures.
Functional scaling means that a network is able to extend both horizontally (more functionality) as well as vertically (different ways of integrating abundant functionality).
New functions must be integrated; otherwise we do not have scaling of functionality, only function accumulation. In the context of ANA, the ability of functional scaling is considered vital, as this provides the basis to allow dynamic integration of new, autonomic functionalities, as they are developed.
FP6-IST-27489 ANA Project - Deliverable D1.4/5/6v1 – ANA Blueprint – First Version Page 1
The Internet and the OSI model are examples of a functionally non-scalable approach:
both are based on the premise of factoring out functionality. The “canonical set” of protocols collected in the Internet-Suite has done a good job so far; however the limitations of this one-size-fits-all approach have become visible. This is not surprising, since the networking layer of the Internet has not aimed at functional scaling by design: it has evolved through a ‘patchwork’ style. Additions were made in a stealth-like way and have not dared to change things. The main examples are the introduction of the hidden routing hierarchy with AS, CIDR or MPLS, as well as other less successful initiatives such as RSVP, multicast, mobile IP and MANET.
As a result, the objective of the ANA project is to provide an architectural framework – a “meta architecture” – that allows the accommodation of and interworking between the full range of networks, ranging from small scale Personal Area Networks, through (Mobile) Ad hoc Networks and special purpose networks such as Sensor Networks, to global scale networks, in particular the Internet.
An important step towards such a “meta architecture” is to design a framework that is able to host different types of network. As a result, ANA encompasses the concept of Compartments as a key abstraction that allows co-existence and interworking of different types of network through a minimum generic interface.
Furthermore, the operation of different types of compartments must be analysed in order to identify the fundamental building blocks of the abstraction. The de-composition of compartments into the core functions helps to understand how the necessary flexibility (functional scalability) to provide autonomicity can be achieved through compartments.
An important concern in developing a new architecture is to ensure that networks have the necessary emergent properties to support the applications and users running on them.
Key amongst these is resilience and security, which we explore in depth in ANA, and evolvability which is an implicit part of the project’s concerns.
1.2 Scope of Deliverable The Blueprint documents focus on the overall architectural aspects of ANA. As such, they define the basic abstractions and building blocks of ANA and present their basic operation and interactions.
As it is most important to establish a sound architectural framework, the focus of the initial work was primarily on the overall architecture. How autonomicity can be achieved will be more carefully studied once the architectural framework becomes stable and the key building blocks and their basic operations have been identified.
Therefore, the blueprint does not capture the detailed results of other areas of work in the project, such as routing, service discovery, functional composition, monitoring, etc. The
outcomes of those areas of work are describes in the following deliverables:
• D2.1: First Draft of Routing Design and Service Discovery 
• D2.2: Functional Composition Framework  FP6-IST-27489 ANA Project - Deliverable D1.4/5/6v1 – ANA Blueprint – First Version Page 2
• D3.1: Monitoring Framework 
• D3.2: Resilience/Security Framework 
1.3 Structure of the document The structure of this document is as follows: Section 2 defines the ANA terminology.
Section 3 describes the core abstractions of the ANA network architecture and the basic communication paradigms. Section 4 further outlines how autonomicity is achieved in ANA. Section 5 defines the ANA node, its basic components and operation as well as the abstract interfaces. Section 6 defines the information flow framework that is the basic concepts for information sharing in ANA. Finally, section 7 concludes the document.
• An Identifier consists of a finite sequence of symbols of a given alphabet.
• Identifiers are used for identification of an entity within a set of entities (e.g., to single out one or more objects from a set of objects).
• Identifiers are typically persistent and globally unique within a given identifier space.
• Within ANA, identifiers are used to identify objects, resources, etc.
• A Name is a globally unique, persistent identifier used for recognition of an entity (e.g., object, resource, and host). For example, a name can be used to resolve the address/locator of the entity.
• Within ANA, names are used to identify an entity (e.g., object, resource, and host). Names can thus be used to resolve the identifier or address/locator of an entity, which is necessary to gain access or communicate with the entity.
• An Address is an identifier that is used for routing and forwarding.
• Addresses entail the information that is needed to transport data/information from a source to a destination.
FP6-IST-27489 ANA Project - Deliverable D1.4/5/6v1 – ANA Blueprint – First Version Page 4
• In case of structured addresses, which define the location of an entity within a topology, addresses are also called Locators.
• A Label is a node-local identifier for an Information Dispatch Point.
Functional Block (FB):
• A FB is an information processing functions in ANA, which for example generates, consumes, processes or forwards information.
• FBs run on one node only, which means a node has full control over the functional blocks it hosts.
• FBs can have zero or more input and output.
Information Channel (IC):