Browsing by Author "Om, Shree"

Now showing 1 - 5 of 5
  • Thumbnail Image
    Item
    Battery and data drain of over-the-top applications on low-end smartphones
    (IIMC International Information Management Corporation, 2018) Om, Shree; Tucker, William David
    Low-end smartphones with sub $50 price tags provide affordable device ownership to low-income populations. However, their limited capacity, when combined with the need for multimodal connectivity, raises usage concerns in rural off-grid regions. Some off-grid regions in sub-Saharan Africa provide recharge facilities using solar power and charge money for the service. Adding data bundle costs to frequent recharge costs, affordability of low-end smartphones becomes questionable in such areas. Community-controlled solar-powered wireless mesh network models with Session Initiation Protocol capability could alleviate the network usage cost conundrum and consume less power in low-end smartphones with the usage of WiFi. This paper reports on investigations that reveal usage of WiFi consumes less battery than 3G, 2G and Bluetooth. In addition, we feel that lowering recharge costs also requires battery consumption knowledge of the over-the-top applications. Using automated voice calls, this paper reports on battery and data consumption by multiple popular social media applications using one type of low-end smartphone. Data consumption was calculated with the objective of learning how to lower data bundle costs by selecting the application with least data consumption. Battery consumption due to CPU usage by the applications was also measured. Results show that WhatsApp consumes the least battery amongst instant messengers and also the least data over all apps measured. SipDroid consumes the least battery overall. Additionally, the reported experiments provide a framework for future experiments aimed at evaluating battery and data consumption by other smartphone applications.
  • Thumbnail Image
    Item
    Evaluating energy consumption on low-end smartphones
    (2015) Om, Shree; Tucker, William David; Rey-Moreno, Carlos; Erasmus, Zenville
    The relationship between battery consumption in smartphones and the usage statistics of a phone is direct. Modern smartphones, even low-end, are equipped with multiple wireless technologies, e.g. GSM, 3G, WiFi and Bluetooth. Each of these technologies has a different energy consumption profile. A wireless mesh project in the Mankosi community in rural South Africa is about to introduce low-end smartphones onto the network. The mesh network is powered with solar-charged batteries because the community at present does not have electricity. Local residents also use these batteries to recharge cell phones at a nominal cost. Introduction of smartphones will increase the recharge frequency as phone usage will increase; thus draining a phone battery more quickly, as well as escalate recharge costs. Thus, the smartphones must be chosen and used effectively in order for batteries to last longer. Related work identifies WiFi wireless technology as the most battery efficient way of transfer when compared to GSM, 3G and Bluetooth. This research proposes experiments to further investigate energy efficiency of WiFi in low-end smartphones that we intend to use for local and breakout voice over Internet protocol (VoIP) calls and data services, on a rural wireless mesh network.
  • Thumbnail Image
    Item
    Investigating battery consumption in low- end smartphones: preliminary results
    (International Information Management Corporation, 2017) Om, Shree; Tucker, William David; Rey-Moreno, Carlos
    Due to lack of electricity in the rural communities of Africa, users often have to incur extra expense on recharging their mobile phones. The limited battery capacities of low-end smartphones, therefore, present a barrier to their positive integration with such communities because of their vast networking features that likely lead to their elevated usage, thus, causing faster battery depletion and in turn, escalating recharge costs. This paper presents preliminary results of ongoing investigations on battery consumptions in low-end smartphones in order to estimate their battery life under different usage scenarios, and estimate the surge in communication expense brought about from the frequent recharging. Voice call experiments over WiFi and 3G data, in different network mode combinations, were conducted using three brands of low-end smartphones. Results, compared using analysis of variance and Tukey methods for pairwise comparison, yielded Smart4Mini brand to be the least battery draining, along with Smart Kicka being second best and Galaxy Pocket Neo being least efficient. In addition, the investigations aided in building a platform for future experiments in order to precisely estimate communication costs under different usage scenarios. Dissemination of such information can assist rural users in making well-informed communication expenditure towards purchase and usage of low-end smartphones.
  • Thumbnail Image
    Item
    Investigation of a dual-band dual radio indoor mesh testbed
    (2018) Tucker, William David; Om, Shree
    Scalability of community wireless mesh networks in terms of mesh nodes and mobile end-users along with their anticipated usage profile remains unexplored territory. However, conducting such experiments on a live community network could cause network breakdown leading to disrupted network services, and discomfort to end-users.
  • Thumbnail Image
    Item
    Towards a scalability model for wireless mesh networks
    (Telkom, 2015) Om, Shree; Rey-Moreno, Carlos; Tucker, William David
    Zenzeleni mesh network is a wireless ad-hoc mesh network that provides voice services using public analogue telephones to the Mankosi community in the Eastern Cape Province. We would like to improve on the network infrastructure by upgrading the mesh routers and introducing low-end smartphones onto the network; and offer both data and voice over Internet protocol services. However, before deploying resources, it is imperative to identify the maximum number of mesh nodes, clients and simultaneous voice over internet protocol calls that can be supported by the mesh network while maintaining acceptable quality of service levels. Absence of such data might lead to financial risk and time depletion when setting up an optimal network. Bolstering the claim are investigations that report drop in quality levels as network density and hop count escalate. As current investigations mostly yield capacity models to predict per-node throughput with increasing hop count, we propose experiments to devise a scalability model to quantify scalability of mesh networks in this paper. We recommend experimental implementations at simulation level in Network Simulator-3 moving on to testbeds built using WiBed, and then finally take results to the field.