The currently trending paradigms of edge and fog computing attempt to provide services close to the end user, to meet the demands of latency-sensitive applications and to limit bandwidth consumption in the network core. One open issue is the pricing of edge and fog resources. Current pricing schemes are usually oligopolistic and not fair. In this work, we propose DMC, a marketplace that can dynamically determine the fair price for arbitrary resource types and instances based on supply and demand existing at that period. Unlike the state-of-the-art solutions, DMC performs integral allocation of resources and thereby avoids the unbounded integrality gap. Additionally, DMC provides differential pricing among instances to allow varying prices based on the perceived value of a resource. We evaluate DMC with both heavy and non-heavy tailed distributions to reflect diverse buying interests and the number of resources sold to demonstrate the feasibility of our solution for several realistic scenarios. We observe that (i) DMC arrives at market-clearing prices; (ii) DMC generates 10x to 100x more profit than state-of-the-art solutions, while still maximizing the Nash Social Welfare to achieve prices that are fair to both buyers and the resource providers; and (iii) the computation time for DMC does not exceed 10 seconds even in the case of 500 resource types with 500 buyers each, making it applicable for real-time use cases.

Currently, cloud computing is a primary enabler of new paradigms such as edge and fog computing. One open issue is the pricing of services or resources. Current pricing schemes are usually oligopolistic and not fair. In this work, we propose OFM, an online learning based marketplace that dynamically determines the price for arbitrary resource types based on supply and demand existing at that period. Unlike state of the art solutions, OFM can handle an arbitrary number of customers and resource types at every instance of time. It further performs integral allocation of resources and thereby avoids the unbounded integrality gap. We evaluate OFM with both real and synthetic datasets to reflect varying buying interests, the number of resources sold and market volatility to demonstrate the feasibility of our solution for several realistic scenarios. We observe that (i) OFM achieves about 9% of optimal prices and maximizes the Nash social welfare (NSW); (ii) OFM converges faster and works with different data distributions; and (iii) OFM scales for a large number of resources and buyers and computational time is in the order of microseconds, making it applicable for real-time use cases especially in edge markets.

In software-defined networking (SDN), high-level abstractions typically offer a useful means to avoid writing network applications and policies on lower levels. However, abstractions are typically developed for a specific use case, which in turn results in an abundance of existing abstractions for different networking tasks. As a consequence, orchestrating these abstractions to implement a common network policy becomes an arduous task. To address this challenge, plain data representations of the network and its control infrastructure have been proposed recently, which offer programmable ad-hoc abstractions to administrators. However, these frameworks suffer from quite complex programming requirements and impractical performance in terms of latency, as they are based on relational database engines. In this paper, we address these shortcomings by introducing Gavel, an SDN controller that at its heart facilitates a plain data representation based on a graph database. By exploiting the native graph support of the database engine, Gavel significantly eases application and policy writing. Additionally, we show by experimental evaluation of several typical applications on multiple different topologies that Gavel offers significant performance improvements over the state-of-the-art solutions.

Virtualization helps to deploy the functionality of expensive and rigid hardware appliances on scalable virtual resources running on commodity servers. However, optimal resource provisioning for non-trivial services is still an open problem. While there have been efforts to answer the questions of when to provision additional resources in a running service, and how many resources are needed, the question of what should be provisioned has not been investigated, in particular, for complex applications or services, which consist of a set of connected components, where each component in turn potentially consists of multiple component instances (e.g., VMs or containers). Each instance of a component can be run in different flavors (i.e., number of cores or amount of memory), while the service constructed by the combination of these component configurations must satisfy the customer Service Level Objective (SLO). In this work, we offer to service providers an answer to the what to deploy question by introducing Rconf, a system that automatically chooses the optimal combination of component instances for non-trivial network services. In particular, we propose an analytical model based on robust queuing theory that is able to accurately model arbitrary components, and develop an algorithm that finds the combination of their instances, such that the overall utilization of the running instances is maximized while meeting SLO requirements.

Today, Online Social Networks (OSNs) are ubiquitous means of communication. In order to prevent the misuse of personal user data by OSN providers, various research efforts have produced a multitude of approaches to decentralize OSNs in the past decade. The most critical challenge for these systems is to replace the infrastructure of centralized OSNs. That is, they need to handle the large amounts of data uploaded by users on one end, and requests towards that data on the other end. Typically, existing approaches instrumentalize cloud facilities or user devices for this task. Unfortunately, they introduce either a monetary cost for users or have limited success in making data highly available. In this work we propose SocialGate, the first prototype that makes use of home routers of users as the infrastructure backbone of the OSN to avoid these shortcomings. Measurements and experiments based on real-world data support the feasibility and practicability of our approach.

SDN (Software-Defined Networking) promises easier management of complex networks. In abiding by this promise northbound APIs play a major role, as they provide abstractions to ease the process of writing network applications. As more and more of such abstractions emerge, one challenging task is to orchestrate different policies that originate from different abstractions. In an effort to provide automation for this process, recent research has suggested to replace the application-specific abstractions by a primitive, SQL-based data representation of the entire network control infrastructure. This representation can then be transformed into higher-level abstractions. However, the use of relational databases results in performance inferior to existing controllers and still requires complex application writing. To address these issues, in this work we introduce Gavel, a graph database based controller which exploits the natural graph structure of computer networks. Based on results obtained from a prototype implementation, we show that Gavel not only simplifies application writing, but at the same time offers a significantly better performance than existing schemes.

Concomitant with the tremendous growth of online social networking (OSN) platforms are increasing concerns from users about their privacy and the protection of their data. As user data management is usually centralized, OSN providers nowadays have the unprecedented privilege to access every user's private data, which makes large-scale privacy leakage at a single site possible. One way to address this issue is to decentralize user data management and replicate user data at individual end-user machines across the OSN.

A Sybil attack can inject many forged identities (called Sybils) to subvert a target system. Because of the severe damage that Sybil attacks can cause to a wide range of networking applications, there has been a proliferation of Sybil defense schemes. Of particular attention are those that explore the online social networks (OSNs) of users in a victim system in different ways. Unfortunately, while effective Sybil defense solutions are urgently needed, it is unclear how effective these OSN-based solutions are under different contexts. For example, all current approaches have focused on a common, classical scenario where it is difficult for an attacker to link Sybils with honest users and create attack edges; however, researchers have found recently that a modern scenario also becomes typical where an attacker can employ simple strategies to obtain many attack edges. In this work we analyze the state of OSN-based Sybil defenses. Our objective is not to design yet another solution, but rather to thoroughly analyze, measure, and compare how well or inadequate the well-known existing OSN-based approaches perform under both the classical scenario and the modern scenario. Although these approaches mostly perform well under the classical scenario, we find that under the modern scenario they are vulnerable to Sybil attacks. As shown in our quantitative analysis, very often a Sybil only needs a handful of attack edges to disguise itself as a benign node, and there is only a limited success in tolerating Sybils. Our study further points to capabilities a new solution must possess; in particular, in defense against Sybils under the modern scenario, we anticipate a new approach that enriches the structure of a social graph with more information about the relations between its users can work more effectively.

With increasing frequency, users raise concerns about data privacy and protection in centralized Online Social Networks (OSNs), in which providers have the unprecedented privilege to access and exploit every user's private data at will. To mitigate these concerns, researchers have suggested to decentralize OSNs and thereby enable users to control and manage access to their data themselves. However, previously proposed decentralization approaches suffer from several drawbacks. To tackle their deficiencies, we introduce the Self-Organized Universe of People (SOUP). In this demonstration, we present a prototype of SOUP and share our experiences from a real-world deployment.

A Sybil attack can inject many forged identities (called Sybils) to subvert a target system. Among various defense approaches, of particular attention are those that explore the online social networks (OSNs) of users in a target system to detect or tolerate Sybil nodes. Albeit different in their working principle, all these approaches assume it is difficult for an attacker to create attack edges to connect Sybils with honest users. However, researchers have found that an attacker can employ simple strategies to obtain many attack edges. In this work we revisit the state-of-the-art, OSN-based Sybil defenses, and point out their strengths and weaknesses due to the impact of the new properties. We find these defense approaches are vulnerable to attackers under the new scenario, and in many cases a Sybil node only needs to obtain a handful of attack edges to disguise itself as a benign node.

Social networking platforms such as Facebook, MySpace, and Twitter have seen a significant increase in user population and user provided information. However, users are increasingly concerned about identity and data privacy since information is controlled by single companies. To address this issue researchers investigated alternative solutions, where the users' data, e.g. profile information, comments and messages, is stored at user-controlled nodes. Although these solutions provide a plausible means for avoiding privacy leaking in central instances, they raise a new challenge to design a cost-effective storage replica scheme which ensures a high data availability even when some users are offline. In this paper we present Gemstone, a social network platform where the data replication scheme leverages a learning mechanism based on social relationships, online patterns of peers and user experiences. Our preliminary evaluation shows that compared to related works, it achieves higher data availability while requiring a smaller number of data replicas.