Celeste is a robust peer-to-peer object store built on top of a distributed hash table (DHT). Celeste is a working system, developed by Sun Microsystems Laboratories. During the development of Celeste, we faced the challenge of complete object deletion, and moreover, of deleting "files" composed of several different objects. This important problem is not solved by merely deleting meta-data, as there are scenarios in which all file contents must be deleted, e.g., due to a court order. Complete file deletion in a realistic peer-to-peer storage system has not been previously dealt with due to the intricacy of the problem - the system may experience high churn rates, nodes may crash or have intermittent connectivity, and the overlay network may become partitioned at times. We present an algorithm that eventually deletes all file content, data and meta-data, in the aforementioned complex scenarios. The algorithm is fully functional and has been successfully integrated into Celeste

Corporate IT as well as individuals show increasing interest in reliable outsourcing of storage infrastructure. Decentralized solutions with their resilience against partial outages are among the most attractive approaches. Irrespective of the form of the relationship, be it based on a contract or on the more flexible cooperative model, the problem of verifying whether someone promising to store one's data actually does so remains to be solved, especially in the presence of multiple replicas. We introduce a lightweight mechanism that allows the data originator or a dedicated verification agent to build up trust in the replica holder by means of protocols that do not require prior trust or key establishment. We show how naive versions of the protocol do not prevent cheating, and then strengthen it by adding means that make it economically attractive to be honest. This provides a foundation for further work in providing trustworthy distributed storage.

One of the solutions to deter copyright violations consists of embedding hard to detect watermarks in digital media. Current developments have focused on how to embed watermarks, and on one-to-one exchanges on how to securely convey tagged data to the end consumer Assuming a large customer base or predistributed media, it may become prohibitively expensive or time consuming to tag each separate copy of data individually before it is delivered to the customer We present two mechanisms that allow the preparation and distribution of tagged data in a more scalable way than existing approaches. The first one, called hierarchical tagging, is preferable for on-line multi-level distribution, where producer and consumer are not in direct contact, but intermediate agents provide distribution channels and sales platforms. The second method is well Suited to pre-produced bulk-media distribution (such as through CDROMs or DVDs), with only small amounts of on-line information being transferred to each consumer We call it bulk-tagging.

Summary form only given, as follows. The past decade has seen a strong opening of company networks towards the Internet. Nearly every organization has some web presence, does some business by email (internally and externally) and many allow their employees access to the Internet from the office. Firewalls (acting as filter and proxy for network traffic) were supposed to be the magic all-encompassing solution to regulate this opening, and not expose the internal infrastructure to the public. But there are problems. The request for transparency and higher accessibility has been getting stronger. Firewalls process more and more traffic, and have to enforce more complex (and harder to formulate) restrictions. They are supposed to offer more and more fimctionality, and they get harder to use all the time. This way, firewalls are becoming a source of faults themselves, and a security risk. P2P Environments reinforce the issues, by potentially opening up many portals between different types of networks. Drive-by hacking in the wireless ethernet world is just one example of this. How do you decide who is going to be a member of your little ad hoc network, and whether users can employ any of the devices participating to hop on (or get routed to) a network they are not supposed to get to? Are there alternatives for classic firewalls? Do they apply to the P2P world? Do they fit the current scenario of ever-increasing mobility and ad hoc intermeshing of our computing environment? The talk explores the rise of firewalls, their evolution and tendencies in this area, and has a look at their strengths and weaknesses. Some alternative solutions are examined, and a vision of a potential future solution is presented.

We present a vision of computing environments in which enterprise networks are built using untrusted public infrastructures. The vision allows for networks to dynamically change depending on the need of their users, rather than forcing the users to build organizations around networks. This vision is realized through a design abstraction called virtual enterprise networking, or short Supernetworking. A first prototype of such a Supernet has been implemented on Linux. Supernetworking introduces a new layer of abstraction in a layered model of computer networking. The Supernet layer sits directly above the network layer and includes its own addressing structure and security services which protect all data transmitted by the network layer. A key component of a Supernet is communications tunneling. Instead of the traditional two endpoints, our tunnels have as many endpoints as there are computers participating in a Supernet. While tunneling has been repeatedly used to implement infrastructure services such as multicasting, virtual private networks, and support for mobility, we distill these technologies into a single, simple abstraction. This new abstraction enables the ability to out-source network infrastructure services in a transparent and secure manner, mobility, and the creation and administration of secure ad-hoc virtual computer networks.

Distributed multimedia applications require a variety of communication services. These services and different application requirements have to be provided and supported within: (1) end-systems in an efficient and integrated manner, combining the precise specification of quality-of-service (QoS) requirements, application interfaces, multicast support, and security features and (2) the network. The Da CaPo++ system presented in this paper provides an efficient end-system middleware for multimedia applications, capable of handling various types of applications in a modular fashion. Application needs and communication demands are specified by values in terms of QoS attributes and functional properties, such as encryption requirements or multicast support. Da CaPo++ automatically configures suitable communication protocols, provides for an efficient runtime support, and offers an easy-to-use, object-oriented application programming interface. While its applicability to real-life applications was shown by prototype implementations, performance evaluations have been carried out yielding practical experiences and numerical results.

Middleware supporting secure applications in a distributed environment faces several challenges. Scalable security in the context of multicasting or broadcasting is especially hard when privacy and authenticity is to be assured to highly dynamic groups where the application allows participants to join and leave at any time. Unicast security is well-known and has widely advanced into production state. But proposals for multicast security solutions that have been published so far are complex, often require trust in network components, or are inefficient. In this paper, we propose a framework of new approaches for achieving scalable security in IP multicasting. Our solutions assure that newly joining members are not able to understand past group traffic and that leaving members may not follow future communication. For versatility, our framework supports a range of closely related schemes for key management, ranging from tightly centralized to fully distributed, and even allows switching between these schemes on-the-fly with low overhead. Operations have low complexity [O(log N) for joins or leaves], thus granting scalability even for very large groups. We also present a novel concurrency-enabling scheme, which was devised for fully distributed key management. In this paper, we discuss the requirements for secure multicasting, present our flexible system, and evaluate its properties based on the existing prototype implementation.

Shared window systems allow collaboration transparent, single-user applications interacted with on multiple users' workstations, enabling the members of a cooperative ensemble to simultaneously share and revise information. The paper presents a system capable of sharing applications running under the X Window System. In contrast to previously implemented systems, the present shared window system addresses issues that are crucial for general-purpose use. The shared window system is policy-free, i.e., there are no preferred policies for handling issues such as admission and floor control. Instead, it offers a set of essential mechanisms on top of which various policies and user paradigms may be realized. Further, the system distributes the sharing functionality among all sites involved in a cooperative activity. Measurements have shown a positive impact of this on the overall performance of the system and thus justified the viability of the design decisions taken.>