The SAIL project focuses on three main technical objectives that address the core of the observed trends. In short, it will provide a comprehensive, working solution for responding to these changes, by integrating a Network of Information and Cloud Networking with Open Connectivity Services.
In more detail:
- The interaction patterns of emerging applications no longer involve simply exchanging data end-to-end. These new patterns are centred on pieces of information, being accessed in a variety of ways. Instead of accessing and manipulating information only via an indirection of servers hosting them, putting information objects themselves at the centre of networking is appealing and has resulted in the notion of a Network of Information. A Network of Information simplifies communication compared to end-to-end or store-and-forward networking, making it more suitable for large-scale, complex applications and for improved programmer and network efficiency. By caching multiple, equal copies of the same information object, a Network of Information natively supports large-scale content distribution. These properties also make a Network of Information suitable for delay tolerant networking. The Network of Information architecture natively supports mobility of information objects, nodes, users, and applications. But it is still unclear how to make sure such a Network of Information-based communication model can be supported efficiently in the network at runtime. For example, the model mandates that information be available near to its point of consumption, but that entails caching inside the network – it is not clear how to integrate such caches in routers, especially as the overall network scales up. However, such an integration will provide substantial benefits, e.g., by reducing information access delay and required bandwidth, improving dependability, and simplify network operation.
- Network applications can fluctuate rapidly in popularity and in terms of the amount of user interaction. This makes provisioning a difficult problem, both on the server and storage side, as well as on the networking one. On the server and storage side, cloud computing has successfully addressed many of these challenges, using virtualisation as a core technique. However, it is still unclear how to provide suitable network support for such highly variable applications when they run not just over the tightly-controlled, custom-tailored network of a cloud computing operator, but inside more complex, more diverse telecommunication operator networks. Even if it is possible to provide the computational resources, it is not obvious how to dynamically provide the necessary networking support/capacity or the complex networking topology required by, e.g., a multi-tier web application. Nor is it obvious how to provide networking support for distributed cloud systems, spanning multiple networks and operated by different entities. A complex example scenario would be a Network of Information not only serving static, but dynamic information (e.g., results of database queries made to travel websites, or cookie-like state of a distributed application), requiring computing resources dynamically allocated and networked in a flash network slice when the Network of Information entities are reaching capacity limits.
- In principle, the raw technologies for efficiently transporting even very large amounts of data are already in place (e.g., optical circuit switching), but these do not mesh well with current Internet control concepts (e.g., packet switching and connection-less networking). Indeed, current solutions struggle with the heterogeneity of the deployed networks and cannot seamlessly integrate control of these resources from the edge (i.e., using the context known to the application). For example, it is practically impossible to exploit diversity (e.g., random variations in channel quality or structural differences in channel properties like different delay/data-rate tradeoffs) existing over different communication technologies between two endpoints, switching between technologies as the flow’s required data rate changes. Similarly, efficient multi-path/multi-protocol/multi-layer optimisation is still infeasible. Cloud Networking, and especially Network of Information, will substantially change the traffic patterns and data flows in the Internet and make the requirement for agile connectivity service even higher.
SAIL deals with particular aspects of a real, complex network. The figure below shows how entities are all reflected differently in Network of Information, Cloud Networking, and Open Connectivity Services, yet all three aspects staying connected and in relation to each other. The figure also highlights that one expects the Network of Information and Cloud Networking aspects to appear multiple times, while the Open Connectivity Services one constitutes a single basis for the other two.
SAIL’s three main technical objectives constitute the three trends that lead to the Network of the Future.
In addition, there will be work as well on various cross-cutting themes (e.g., management, security, inter-provider and prototyping issues) that affect each of the broad technical areas, and will also ensure that standardisation and migration, socio-economics, business and regulatory aspects, and exploitation and dissemination activities are properly considered.
