Technical Reports

Background and Justification:[Back to table of contents]

Providing power to relays and access points (AP), in particular wireless ones which are often located in geographically dominating (and thereby often harder to access) locations, is an ongoing challenge faced by most Internet service providers operating in rural areas in Asia and elsewhere. Chosen relay locations are typified by total absence of grid power supply or, when power is available, frequent power cuts.  As a result, operators install power backup (chargers+ batteries) as well as, in some cases, power-generation (solar, wind, etc.) devices for these relays at a considerable cost. Installation of a power backup system requires “power budgeting” – i.e. ensuring that a specific relay can stay up for a sufficient amount of hours based on its own battery power in the absence of charge – is carried out based on estimates of power consumption of devices installed. These estimates are often based on static/average power consumption data of devices provided by equipment manufacturers. Given the high cost of power backup and generation, ISPs are constantly trying minimize the cost of a power-backup array (by not over-provisioning a site in terms of power generation and batteries) to balance these costs with the considerable costs and customer dissatisfaction caused by downtimes (in case of, e.g. site under-provisioning of power backup). The ability to measure and monitor the actual power generation and consumption is therefore of great operational and economic importance for rural ISP. This “balancing act” task is not trivial – main grid power may be available only at certain times of the day, and fluctuate and autonomous power generation (solar, wind) tends to fluctuate with ambient conditions. Further, the consumption pattern of relays changes with the quantity and type of equipment used, the duration of transmission and reception activity of various devices, also impact the discharge intensity (the higher the activity and number of devices, the faster the discharge per given array). Previous attempts? While some devices and systems for addressing this issue exist, they often have only limited applicability for rural ISPs. Some do not deal with relatively low power setups typical for wireless relays, and many tend to be quite costly and therefore are not usually used by cost-sensitive rural operators. In addition, many of these devices are not equipped with remote/IP-based reporting capabilities, thereby further reducing their usability for operators whose relays and devices are located across wide terrains (and are, by default, all connected to the internet therefore making the task of linking them to a central data collection and repository relatively easy). Motivation of your organization In the absence of connected power monitoring systems, operators such as AirJaldi relay on “guesstimates” – IF they know of a disruption to power supply, they will estimate the approximate time a relay can keep operating without power and try to tailor their response accordingly (e.g. replacing depleted batteries in cases of long downtimes leading to a total depletion of a relay’s power array). This results often in costly miscalculations – for example when a relay goes down earlier then assume or, in an opposite case, a team is sent too early to address the problem (e.g. with fresh, charged batteries) only to find that site’s power provision is active again. As noted earlier,  addressing the chance of such mistakes by over-provisioning of equipment, may lead to unnecessary costs. Our project sought to address these constraints by developing an affordable cloud-based power monitoring and alert system that will be able to measure and report on the actual power charging and consumption patterns of each individual relay, predict the approximate occurrence of a downtime  and alert the operator, allowing for adequate response time and power budgeting. Theory of change The ultimate outcome of the project is for ISPs and their customers (e.g. power-dependent sensor arrays) to be able to ensure a constant, uninterrupted service at an affordable cost. In due course, the system developed here could be used to provide additional information on the activity and health of connectivity and related devices installed on relays or within networks. This is achieved through outputs provided by sensors connected to the battery array, reporting on the battery’s charge status. These readings, loaded into a cloud-based repository, are used to create a relay-specific power generation and usage patterns, enabling more precise prediction of the time of downtime of a specific relay, thereby providing sufficient (but not excessive or unnecessary) time -to-reaction and, more precise power budgeting for each relay.

Project Implementation:[Back to table of contents]

Note: as this section contains images, the version with images is available here.

Project Evaluation:[Back to table of contents]

• • To what extend the project achieved its objectives? What contribution to development did the project make? Is there evidence of positive impact?

We were successful in developing a working system and UI enabling users easy access to the data and, most importantly, enabling ISP oprators to take more informed decision on the way they set up and operate the power provision and backup arrays on their relays. As the project has just been completed, we do not yet have evidence of any impact, but will be able to provide information on this at a later stage.

• • To what extent the project delivered benefits/outcomes that support gender equality, diversity and inclusion?

As noted elsewhere here, the project has not explicit gender and diversity component. It is howver worth noting, in this context, that our work is mostly based in rural areas and that many of the areas where we work have a diverse ethnic and religious populations.

• • To what extent has the project lived to its potential for growth/further development?

In our opinion the project has achieved its objective, in particular the first three, which deal with building the systems. The potential for further growth and development surely exists, mainly in integration of machine learning components and in addition of similarly affordable systems for measurement of relay health and performance.

• • To what extent do the proposed activities support the development of local technical capacity in a gender equality and inclusion framework?

The project contributed to the development of the skills of our own technical team as well as external local entities we harnessed to help us in the project. The finished product helps local teams address operational needs and challenges and hence strengthens the quality and efficiency of their work and their ability to provide better service to our customers.

• • What were the most important findings and outputs of the project? What will be done with them?

Apart from the inherent benefits that the TellTale system will bring to us and other field operators, the project has shown us that addressing some of our technical challenges can be a task we could address with our own strengths. Further, our need to constantly take into account financial realities of rural operators and our deep and ongoing knowledge of rural operations mean that we are in a good position to design, develop and implement technical solutions that are relevant to the realities of rural operators. We  are sure this project will lead us to developing additional, similar, solutions.

• • What lessons can be derived that would be useful in improving future performance?

First, we should include components of machine learning (part of the planned project which we were not able to implement) to further improve the prediction of the model through dynamic analysis of each relay’s actual performance. Second, the addition of battery temperature measurement will enable better understanding of battery health (in particular charging regimes).  This is of particular importance when more expensive batteries (e.g. Lithium-Ion) are used.

• • To what extend the project help build up the capacity of your institution or of the individuals involved?

As also noted earlier, the project has enabled our technical team to engage in domains that they may have not necessarily encountered as part of their routine work. In doing so, the team, apart from developing project-specific work, has strengthened its capacity – and importantly, belief – to undertake R&D projects that are related to AirJaldi’s work on rural connectivity and connectivity-related applications.

• • Were certain aspects of project design, management and implementation particularly important to the degree of success of the project?

We view the following project aspects as being of prime importance:

• The emphasis on cost.
• Our focus on what is essentially important for making the projections, rather than simply collecting a whole range of data inputs.
• Reliance on cloud analysis, which opens the door for integration of additional sensor components and for integration of the system within other network monitoring and management systems (e.g. geotagging of relays and customers and measurement of the impact of that on performance, etc.).

Gender Equality and Inclusion:[Back to table of contents]

The project had no explicit gender component. Its contribution to gender equality is indirect, mostly through improvement of the quality and stability of ISP operations.

Project Communication Strategy:[Back to table of contents]

In the TellTale project, the main target community for dissemination are rural ISPs and affiliated groups (researchers, students, companies operating in the connectivity space, activists, etc.). Our dissemination efforts will be directed at this audience, and includes the following: First, informing people about the project - we will publicize the project and the products via our website, social networks (LinkedIn, FB, twitter, etc.) and will reach out to organizations and online platforms. Second, making the product available – the manuals, codes, etc. are available at a repository we created on GitHub. In addition, the Android app is available on the Google Play store (its presence and overview may help other potential users learn about TellTale) Third – we will be offering support services for those implementing the TellTale system. Support will range from free advice to a “freemium” model where we will help interested parties in one or more of the following: making the Telltale field device, linking and managing the cloud analysis and reporting systems and full operation of the service (making, installing, configuring, managing). The interaction with people and groups through these three channels of dissemination will help us to gain an intimate understanding of the interest in this product, the rate and success of implementation of TellTale outside of AirJaldi, and the impact the product has on the quality of service and ultimately – the productive use of internet connectivity (which will surely benefit from lower downtime occurrences).

Recommendations and Use of Findings:[Back to table of contents]

Recommendation to other practitioners:

• Do not hesitate to ask questions. Asking the right questions is often the first step to getting an answer, whether from existing sources (try forums like stackoverlow, Arduino forums, etc.), or otherwise from within your own or like-minded entities.
• Try to do work in-house, solutions available in the market might not fit your needs, needs which after all you are most familiar with. We always knew the monitoring of battery health is important but were stuck with trying different solutions which did not effectively addressed our needs and challenges.
• Don’t invent, innovate. Huge amounts of information and solutions are already available out there, and many can be used free of charge. Your task, if you are (as we are) a field-based, production-oriented entity is to put together a package that responds to your needs, is scalable in terms of number of implementations, and  is economically-viable.
• Be flexible – the way we conduct our operations changes all the time. Rather then locking yourself into a rigid solution platform (hardware and or software, cloud hosting and analysis, etc.) try to make us of products (created by you or others) that lend themselves to modification and expansion and, if they become obsolete, will not be too costly to replace.

For ISIF secretariat: The grant helped us in setting a testbed and develop a product that, without the support and push, we may have never attempted to develop. The secretariat has also exercised exceptional patience with us as we were late in completing the project on the date prescribed in the project agreement. We are grateful for both and hope that despite this delay, ISIF, us and the community of broadband network operators will benefit from the fruits of this project.

Bibliography:[Back to table of contents]

https://forum.arduino.cc – forum is very helpful and responsive. https://stackoverflow.com – any useful posts and questions can be asked. https://ohmslawcalculator.com/voltage-divider-calculator - Explanation on how voltage divider works. https://www.allaboutcircuits.com/tools/voltage-divider-calculator - Another explanation on voltage divider. https://www.electronicshub.org/interfacing-voltage-sensor-with-arduino - An article with code on voltage sensing using Arduino.