Intelligent preparation and use of data for more proactive network management

The multiplicity of RMS systems, coupled with communication errors presents a major challenge to MNOs and towercos looking to obtain a “single source of truth” regarding their operations. TowerXchange speak to Infozech Founder and CEO, Ankur Lal to discuss how the company is working with customers to better handle raw data and analyse it in a structured way to equips MNO and towercos with the tools to take action on their sites.

TowerXchange: An increasing number of RMS platforms, either as standalone systems or built into different equipment on cells sites gives towercos and operators a wealth of data with which to monitor operations. Can you explain some of the challenges that are presented to tower owners and operators by this? 

RMS systems are increasingly be deployed by towercos and MNOs with a view to increase site visibility. This need is especially large with new sites (where RMS comes built in) or retrofitting existing sites. In the retrofit scenario, one emphasis is on gaining better visibility of source of power and visibility of fuel consumption.

Towercos believe that once they have the necessary RMS deployed, they will get authentic information and achieve a single source of truth. On the ground there are multiple systems, some standalone and others which are integrated. Due to multiplicity of systems and ground realities often “single source of truth” through RMS deployments is not achieved.

Each RMS system produces data in its proprietary protocol. There are multiple challenges being faced by the towercos with an increasing number of RMS platforms and collection of data from RMS.

The communication between site and server consists of four steps:

1. RMS controller capture and relay: RMS controller captures information at site and relays it.

2. SIM/ and mobile modem transfers information from site

3. Mobile network establishes connectivity between site and server,

4. Server: Set up to receive information from RMS controller and auto correct any transmission errors

In the event that any of the above four steps fail to work, or do not work in tandem, the desired outcome of receiving data on server is not achieved. Whilst this seems simple and should always work, in reality, due to multiple choices for each of the above, getting all of them to work reliably needs focus and review.

TowerXchange: Please can you explain a case where customers had experienced challenges in this area?

Our customer experienced a case where the RMS was configured to relay energy data every ten minutes, meaning six times an hour or 144 times a day, as such, the expected number of data packets for energy data was 144 per day. When we looked at this analysis across sites we found a large number of sites doing this, however for over 30% of the sites, the number of data packets received was under 135 (i.e. they were not 95% compliant). When we delved further we realised that this was happening due to one of the aforementioned reasons, ensuring you are getting all the data is absolutely critical.

While looking at this we also saw sites which were sending 155 packets or more a day – We found that some of the data is erroneous and did not provide a consistent trend, rather it led to confusion of the receiving system. We explained this phenomenon as noise which was due to unpredictable events at site or in transmission of data and as such it resulted in us building sophisticated proprietary solutions to parse the data in such a way, that noise reduction happens.

TowerXchange: Can you share the work that Infozech is doing to better prepare the raw data that is being received by towercos and MNOs on their sites? What further steps are required by different stakeholders to assist in this?

The raw data packets being received by the server may contain alarm information, data value or both. Some of the functionality which helps us in better handling of raw data are:

1. Handling alarm fluctuations: Alarms are configured for key events such as door open, low fuel, low voltage, fire et cetera. Some of them need immediate action while others are not so critical.  At times, the information received may not be fully accurate,  for instance an alarm fluctuation due to a malfunctioning of a sensor or controller can result in the central server starting to receive a flood of these alarm packets. If an alert is passed on every time such event occurs, the end user mailbox is flooded with messages and it becomes difficult for the user to handle.  The Infozech itower (Tower Product suite) comes built in with “intelligent filtering” to assess if this event is due to malfunctioning at site, and sends only the relevant notifications to the user.

2. Fuel sensor calibration: Each site has different types of fuel tanks (of all sizes and shapes) and there are multiple types of fuel sensors which can be fitted. While some newer ones like Capacitive maybe more reliable, others are less so. RMS vendors need to calibrate the sensor with the tank and the equipment to ensure the readings are accurate. In the event that a sensor or tank is changed, recalibration is required.

Besides recalibration we noticed that even the best of sensors have fluctuations due to external temperature and other factors. At times, this fluctuation is so significant that it may cause the receiving system to misinterpret the information; in case of fuel, such as a normal fluctuation may be misinterpreted as theft or vice versa. Infozech has developed a proprietary “Smoothing Algorithm” which helps normalise the data and show the correct trends without fluctuations.

3. Missing data – correction: At times the data may be missing in a data stream. Infozech has a mechanism to identify and highlight any missing data parameters in the raw packet. Infozech has built proprietary algorithms to “fill-in” for missing data depending on the type of data missing and the number of instances in which it is missing. This helps autocorrect a data stream.

4. Business rule: When raw data is received, it needs to be verified against permissible ranges and likely values. Infozech’s system validates every raw data packet and each value in the data packet against its defined type and permissible value and filters out any garbage value from the system. It is possible to configure multiple types of business rules to identify data and depict business scenarios.

Based on these findings, towercos and the RMS vendors should acknowledge and rectify these data discrepancies highlighted by the system in a time bound activity; any loss of data packet due to rejection of garbage values will lead to loss of information and indirectly hamper other day to day processes linked with those data values.

TowerXchange: In the instance where there are multiple images of the same data from the different systems, or where there is a gap in data, what are Infozech’s recommendations on how to best manage this?

The ultimate objective is to treat RMS data as a single source. However, the challenges enumerated above while collecting and validating data leads to compromisation of the objective. This happens because towercos are collecting data from multiple sources.

Infozech recommends the use a standard platform which can reconcile data from any sources based on:

- Type of data (instantaneous value or cumulative value)

- Business rules applied on the data

Infozech’s platform has been precisely designed to achieve this objective for towercos. Infozech provides a platform where it reconciles data from any source or platform. The system automatically eliminates repetitive inflow of the same data, selecting the best value and capturing it. Users can then have an option to approve the best value based on their business requirement for future use.

TowerXchange: With the number of parameters that can be measured on a cell site being seemingly endless, what metrics do Infozech think are particularly important for tower owners to measure that may not be widely monitored currently?

Key things which can be implemented as energy/cost saving measures include:

1. Where there is high genset run hours, battery backup hours should be monitored

2. Where there is high grid availability, generators can be removed by enhancing the battery bank

3. Second (spare) gensets can be removed from sites with lesser load by managing one genset with proper battery cyclic operation

4. Periodic analysis of runtime distribution across the grid, battery and genset based on load and battery capacity which can then lead to up gradation if required.

5. Maintenance results of battery (e.g. discharge test) periodically can suggest enhancement or replacement of battery bank.

6. CPH establishment based on site category based on load, temperature, colocation and other known/unknown factors. This can be then improved based on a feedback from the system correcting variances between actual and theoretical values.

Infozech’s i-Analytics solution helps customers carry out such analysis easily and repeatedly thus helping them take much more informed and optimal decisions.

TowerXchange: Preparing the data into a manageable format is the first step but turning data into intelligence that can be used by the client is key. Where are we today in being able to consolidate and analyse all the different inputs into real intelligence? How do Infozech see this being built upon in the short, medium and long term?

“Improving Profitability through ‘Discipline of Action” (TowerXchange Issue 17, August 2016), focuses on how to assist customers take action. Taking action is often associated with higher risk or effort. Infozech’s iTower platform assists operators capture and analyse information is a structured way.

Once actions are taken, they are fed back in the system for further assessment of the quality of action. This helps assess action effectiveness.

The analysis provides trends which can indicate short, medium and long term actions. One area which Infozech has started engaging in, is the optimal mix of capex – opex. Often a large number of capex measures, such as long lasting batteries, need effective systems which can monitor and measure energy spend and battery life over multiple years to determine:

1. Whether the initiative itself was right (i.e. switching to long lasting batteries), or was it fraught with failures, site downtime issues or difficulty in maintenance.

2. In case the initiative was a success which battery provider gave the best service – in which case was the yield the most – this may not have been in the lowest cost one. In absence of such analysis companies lean towards the lowest cost alternative which at times could even be the highest cost one.

TowerXchange: Can you share some examples where Infozech has worked with a client to get more out of the data that they are generating? What improved efficiencies, cost savings or timelines has this afforded the client?

Infozech has been working with one of our clients to help them to get more value from the data which they receive from RMS systems for cross functional consumption.

One of the challenge faced by the client was the ability to obtain accurate and complete data for energy billing from the RMS system. Major challenges included:

1. Missing /Garbage data packets

2. Incomplete data packets

Incomplete data packets: The Infozech System derives the possible value based on trends in the historic data. There are sites where the data is not available for the complete month, for instance, energy billing cannot be done for site where data is only available for 25 days. In such scenarios of incomplete data, Infozech help in determining those values based on Infozech’s proprietary algorithm “Fill in”.

After applying the above functionality, the customer is now able to bill the sites accurately based on the RMS data.

We have also been providing solutions where data has been missing for 15 days or there is no data for the month. These functionalities can be reapplied to fuel sensor data, genset run hours and other data sets.

---

Infozech will be hosting a roundtable on the subject of “Intelligent preparation and use of data for more proactive network management” at the TowerXchange Meetup Africa & Middle East, being held on 19-20 October at the Sandton Convention Centre. Click here for more information.

---