Article Preview
Top1. Introduction
The continuous evolution in the area of mobile computing has resulted in the pervasive presence of mobile devices all around us. A vast majority of population owns smart phones or a range of other mobile information devices. Along with the surge in numbers, the computational and storage capabilities of mobile devices have also witnessed a significant improvement over the years. A smartphone of the present day is a strong competitor to a personal computer in terms of RAM, CPU, storage, internet accessibility etc. Further, a collection of nodes located in proximity is capable of performance much higher than an individual or isolated node. This motivates the use of mobile devices beyond their conventional use for communication purposes. Consequently, these devices are increasingly being used for computation as well as storage needs of users.
Mobile ad hoc networks (MANETs) and mobile peer to peer networks are examples of networks involving only mobile nodes which have been used for multiple purposes (Darby III & Tzeng, 2010; Wang, Chen, & Gao, 2006). This paper presents a distributed mobile storage system based solely on mobile nodes and ad hoc network connections between them. The objective is to utilize the unused storage capacity of mobile devices located in vicinity to provide transient storage capability to their users which is larger, more secure and reliable than capabilities of their individual devices. However, utilizing only mobile nodes for devising a storage system presents numerous challenges related to loss of data, inconsistency of stored data due to random movement of nodes and threats to the privacy and security of users. The proposed protocol deals with these challenges to develop a distributed storage system using only mobile devices and without any dependence on infrastructure or internet connectivity.
The proposed system will be immensely beneficial when people with similar interests convene in a localized geographical region for a short duration, and need to share, store and access common data with each other. For example, a team working on a project needs to share project data among each other or people gathered together for a conference have to share meeting notes or other important files. The existing solutions include sharing data via Bluetooth, emails or accessing data kept on the cloud. This means a heavy dependence in the internet for the purpose of sharing data. Besides using Bluetooth and emails repeatedly can be very cumbersome and time consuming. We propose a storage system that uses the mobiles of people to store common data in the network as a whole. In this way free or idle memory of mobiles can be utilized to create a temporary storage system. This system can be a more viable alternative than the existing cloud-based storage solutions in situations where cloud services are inaccessible due to absence of or overloading of Wi-Fi or cellular network connections (Cai, Xu, Jiang, & Vasilakos, 2017). In such scenarios, the proposed system can provide a solution for the problem of data storage as well as data sharing.
The proposed architecture uses mobile phones of the users connected over an ad hoc network for storing their data and sharing it with other users. The storage is based on an erasure coding scheme, whereby the data is encoded into data and parity blocks prior to storage. The resultant encoded blocks are then distributed amongst the constituent nodes of the MANET. The encoding results in redundancy and therefore increases the reliability of the system without incurring the overhead of replication. Moreover, this pays attention to the security of the stored data since the complete data is not present with any single node.
For instance, in a meeting, if group members share some common files that are important for all members, it results in an individual copy of each such file to be stored at every mobile node in the network. However, the proposed protocol will reduce the storage requirements per file and further provide a two-fold benefit to the users. Firstly, it will allow the facility of transient storage and easy access to the common files. Secondly, it will remove the dependency on any external storage such as cloud-based storage or the need of any fixed infrastructure in the network. Thus, the mobile nodes will be able to serve the storage requirements on their own and the stored files will be accessible by all the nodes in that network.
The contributions of the work are: