Transmission failures hounded the 2010 and 2013 elections.
In the 2010 elections, about 8 percent of some 78,000 PCOS machines failed to transmit election returns to their respective destinations. At an average of 600 voters who voted, this rate represents 3,744,000 ballots.
It was worse in the 2013 elections. At least 23 percent of about 78,000 PCOS machines or about 17,940 PCOS machines failed to transmit election returns. At an average voter turnout rate of 600 voters, the number of machines that failed to transmit represents 10,764,000 ballots!
Big numbers! It is unknown what happened to the untransmitted election returns in 2010. But for the 2013 senatorial contest, the Commission on Elections (Comelec) came out with what it called “Grouped Canvass Results” as basis for the proclamation of winning senatorial candidates
The official count transmission route is from the PCOS at the clustered precinct to the city or municipal canvassing and consolidation system (CCS) servers where the election returns from the clustered precincts are consolidated and a certificates of canvass is generated. City and municipal candidates for mayor, vice mayor, and councillors are proclaimed at this level while the certificates of canvass for the district, provincial, and national posts are transmitted to the provincial CCS servers. The certificates of canvass from the city or municipality are consolidated by the provincial CCS servers and the provincial certificates of canvass are generated. Provincial candidates for governor, vice governor, members of the Sanguniang Panlalawigan, and members of House of Representatives are proclaimed at this level while the certificates of canvass for national candidates are transmitted to the national CCS servers. In both the 2010 and 2013 elections, the provincial certificates of canvass for the senatorial and party list contests were transmitted to the Comelec national CCS server for consolidation. The Comelec proclaims the winners of the senatorial and party list contests. In the 2010 elections, the provincial certificates of canvass for President and Vice President were transmitted to the national CCS server in Congress for consolidation. It is the duty of the Senate and House of Representatives, jointly convened, to proclaim the winner of the Presidential and Vice Presidential contests.
Election returns also take an unofficial count transmission route. By law, the election returns from the PCOS are supposed to be transmitted to the servers of the majority party, the dominant minority party, the Kapisanan ng mga Brodkasters ng Pilipinas (KBP), and the accredited citizens arm. As implemented, however, the election returns were transmitted to the Comelec central server and what is referred to as the transparency server. The servers of four parties identified by law to directly receive the election returns are connected to the transparency server. The web server is supposed to receive the election returns from the Comelec central server.
It is good practice in systems design to have receiving devices or networks acknowledge data received from source devices or networks to ensure completeness of the data. It is also good practice to provide a system of synchronizing the data received by several devices or networks from the same source to ensure data integrity.
As implemented, however, best practice has been ignored. Data transmission from the PCOS is one way. The receiving city or municipal canvassing and consolidation system (CCS) server was not designed or programmed to send an acknowledgement message to the source PCOS. A system of synchronization check between and among the central server, transparency server, and the city or municipal CCS servers was not provided. The PCOS was designed and programmed to simply try and transmit the election returns to its target destinations in a round-robin manner, that is, if the PCOS senses that it is unable to transmit the election returns to a target, the PCOS will try to transmit to the next target. It will continue to do so until the period of transmission lapses. There was no way for anyone to determine if all transmission targets have the same data. The incompleteness of election returns in the Comelec data may have been a result of the absence of a system of data synchronization.
The failed transmission rates of about 8 percent in 2010 and 23 percent in 2013 are culled from the transparency server only. Data from the central server and the city or municipal CCS servers have not been made available.
There are also other considerations in designing a transmission network. Among these are (1) availability of infrastructure, (2) the terrain where the voting centers are located, and (3) quality and positioning of transmission devices. The automated election system uses primarily the wireless networks of the telecommunications providers. Signal strength depends on the distance of the voting center from the nearest tower of the telecommunications facility.
Let’s face IT. To ensure credibility of election results, the Comelec must resolve the transmission woes experienced in the past two elections. It must work closely with telecommunications providers. Where there is a deficiency or absence of coverage, alternative solutions may be considered for implementation.