P2P storage systems: Study of different placement policies

Abstract

In a P2P storage system using erasure codes, a data block is encoded in many redundancy fragments. These fragments are then sent to distinct peers of the network. In this work, we study the impact of different placement policies of these fragments on the performance of storage systems. Several practical factors (easier control, software reuse, latency) tend to favor data placement strategies that preserve some degree of locality. We compare three policies: two of them are local, in which the data are stored in logical neighbors, and the other one, global, in which the data are spread randomly in the whole system. We focus on the study of the probability to lose a data block and the bandwidth consumption to maintain such redundancy. We use simulations to show that, without resource constraints, the average values are the same no matter which placement policy is used. However, the variations in the use of bandwidth are much more bursty under the local policies. When the bandwidth is limited, these bursty variations induce longer maintenance time and henceforth a higher risk of data loss. We then show that a suitable degree of locality could be introduced in order to combine the efficiency of the global policy with the practical advantages of a local placement. Additionally, we propose a new external reconstruction strategy that greatly improves the performance of local placement strategies. Finally, we give analytical methods to estimate the mean time to the occurrence of data loss for the three policies.

Peer-to-Peer-Based Social Networks: A Comprehensive Survey

“…Irrespective of the overlay upon which the storage technique has been designed on, the most important consideration is guarantee on data availability. In most distributed systems, this is achieved through the inclusion of data redundancy mechanisms which utilize replication and/ or erasure codes [128]. We discuss how this is achieved.…”

“…In addition to selecting an appropriate coding technique, another challenge is maintaining a minimum number of data fragments in the network for durable long-term storage in spite of failures by ensuring proper fragment placement. It has been shown that the choice of fragment placement has an impact on system performance [128]. Therefore, not only is the coding technique important, but also the replica placement policy.…”

“…Therefore, not only is the coding technique important, but also the replica placement policy. Examples of placement policies include SN Computer Science global & random policy, chain policy and Buddy (or RAID) policy [128].…”

An innovative RegTech approach to financial risk monitoring and supervisory reporting

“…In this paper, we propose a novel RegTech[1] approach for financial transaction as well as financial risk reporting, which is based on cutting-edge distributed computing and decentralised data management technologies such as distributed ledger (DL) (Swanson, 2015), distributed storage[2] (Chandra et al, 2013;Caron et al, 2014), algorithmic financial contract standards (Brammertz and Mendelowitz, 2014;Breymann and Mendelowitz, 2015;Braswell, 2016), automated legal text [3] (Hazard and Haapio, 2017) and document engineering methods and techniques [4] (Glushko and McGrath, 2005). Our approach also takes inspiration from the concept of the "bearer service" and its capacity to span over existing and future technological systems and substrates (Kavassalis et al, 2000, Clark, 1988 [5].…”

Objective assessment of the performance of data grid replication strategies based on distribution quality

“…In recent years, more and more works focused on the replica placement in data grids (Benoit et al, 2008;Grace and and Manimegalai, 2014;Hamrouni et al, 2015b). This problem was considered as a complex optimisation problem (Challal and Tebibel, 2010), and several strategies and heuristics are proposed in the literature in order to solve it (Caron et al, 2007;Rasool et al, 2007). Indeed, there is no polynomial time algorithm for the problem that yields to an optimal solution, i.e., an optimal placement of replicas will always lead to exponential time algorithms (Thara et al, 2013).…”