SkyEye: A tree-based peer-to-peer monitoring approach
Section snippets
Motivation
Peer-to-Peer (p2p) networks are well-studied decentralized structures which operate in a self-organized fashion, but also without any global overview. However, it is very useful to keep an eye on metrics or key performance indicators of a running p2p system, e.g. in order to identify weaknesses, bugs and misleading trends. The basic prerequisite of adapting a dynamic p2p network is a precise and efficient monitoring mechanism for collecting and disseminating global information emergence.
Related work and design decisions
Having discussed the requirements of a monitoring solution, in this section, we classify related work by their design decisions and outline the differences to our solution SkyEye. Having a view on the global statistics of a network allows to consider the information on the quality and dynamics of the network for optimized decisions. Monitoring information is used in [3] to foresee low availability of replicating nodes and nodes with relevant duties and thus to counteract by selecting further
SkyEye: Establishing the core monitoring topology
Next, we describe the idea, model and creation of the topology of SkyEye on top of a kbr-compliant [1] overlay. In Section 4, we describe then how this SkyEye topology is used to gather and disseminate the monitoring information.
The main idea for setting up the SkyEye topology is that each node computes based on its overlay ID and its responsibility range its position in the monitoring tree. The (inner) nodes in the monitoring tree are responsible for specific IDs, so called Domain Keys, which
SkyEye: Information aggregation and monitoring protocols
Monitoring statistics give insights into the quality characteristics of a running p2p system. System-specific characteristics are described in a set of metrics, which are monitored and measured permanently on each peer. We use the term metric both for the classification and type of a measured value (e.g. hop count) and also for the actual value (e.g. 5). In this section, we describe the aggregation of metrics and the protocols and mechanisms used to gather, aggregate and disseminate
Evaluation
In this section we evaluate the structured SkyEye, which communicates using a newly established tree topology on top of a DHT, with Push-Sum, a gossip-based approach by Blasa et al. [2]. We outline, compare and discuss similarities and differences caused by their structure in a parameter impact study in terms of its accuracy, adaption and costs in scenarios with and without churn.
Conclusions and future work
SkyEye implements a monitoring layer for structured p2p systems providing the API for key-based routing. It creates an additional (unified) ID space, which is recursively partitioned in ID intervals called Domains. For each Domain, a characteristic ID is calculated using a deterministic function that maps the Domain to a single ID in it, called Domain Key. The peer responsible for the Domain, termed Coordinator of the Domain, is identified by being responsible for the Domain Key in the DHT. The
Acknowledgments
This work was supported by the DFG Grant GR4498 1-1 (OverlayMeter).
References (33)
- et al.
T-Man: Gossip-based fast overlay topology construction
Comput. Netw.
(2009) - et al.
White space: Definitional perspectives and their role in exploiting spectrum opportunities
Telecommun. Policy
(2016) - et al.
Towards a Common API for Structured Peer-to-Peer Overlays
- et al.
Symmetric Push-Sum Protocol for Decentralised Aggregation
- et al.
Towards a P2P Cloud: Reliable Resource Reservations in Unreliable P2P Systems
- et al.
Overlay Bandwidth Management: Scheduling and Active Queue Management of Overlay Flows
- et al.
DiDuSoNet: A P2P Architecture for distributed Dunbar-based Social Networks
Peer-to-Peer Netw. Appl.
(2016) - et al.
Ca-Re-Chord: A Churn Resistant Self-Stabilizing Chord Overlay Network
- et al.
Monitoring and Management of Structured Peer-to-Peer Systems
Monitoring and Management of Peer-to-Peer Systems
(2010)
Some principles for quality of service management
Distrib. Syst. Eng.
SOMO: Self-Organized Metadata Overlay for Resource Management in P2P DHT
Aggregating Information in Peer-to-Peer Systems for Improved Join and Leave
Willow: DHT, Aggregation, and Publish/Subscribe in one Protocol
Cited by (8)
To improve scalability with Boolean matrix using efficient gossip failure detection and consensus algorithm for PeerSim simulator in IoT environment
2022, International Journal of Information Technology (Singapore)Towards Self-Adaptive Peer-To-Peer Monitoring for Fog Environments
2022, Proceedings - 17th Symposium on Software Engineering for Adaptive and Self-Managing Systems, SEAMS 2022LibreSocial: A peer-to-peer framework for online social networks
2021, Concurrency and Computation: Practice and ExperiencePeer-to-Peer-Based Social Networks: A Comprehensive Survey
2020, SN Computer Science