Andhra Pradesh Rolls Out an Elec­tronic Pub­lic Dis­tri­bu­tion Sys­tem Us­ing Open Source

Totally based on bio­met­ric au­then­ti­ca­tion, the elec­tronic pub­lic dis­tri­bu­tion sys­tem project ini­ti­ated and im­ple­mented by the state gov­ern­ment of Andhra Pradesh has been de­vel­oped on Post­greSQL, the open source data­base.

OpenSource For You - - Contents - By: Jag­meet Singh The au­thor is an as­sis­tant edi­tor at EFY.

It was in 2014 that the Andhra Pradesh gov­ern­ment de­cided to go dig­i­tal and opted for an elec­tronic so­lu­tion for its PDS (pub­lic dis­tri­bu­tion sys­tem). While the ini­tial model was tested in a few small re­gions, the fi­nal re­lease can han­dle as many as four mil­lion trans­ac­tions per day dur­ing peak time across all the 13 dis­tricts in the state—all this pow­ered by Post­greSQL, the open source data­base.

The Andhra Pradesh (AP) gov­ern­ment gave the task of trans­form­ing its pub­lic dis­tri­bu­tion sys­tem to a team at the Na­tional In­for­mat­ics Cen­tre (NIC). Led by civil sup­plies commissioner and state in­for­mat­ics of­fi­cer (SIO) of Andhra Pradesh, K. Ra­jasekhar, the NIC team pre­ferred US-based

En­ter­priseDB to build a pro­gres­sive ePDS (elec­tronic pub­lic dis­tri­bu­tion sys­tem) model, in or­der to im­prove the trans­parency, ef­fi­ciency and ef­fec­tive­ness of the sys­tem. The ar­rival of the ePDS in the re­gion has also set the stage for next-gen­er­a­tion so­lu­tions, in­clud­ing eSCM (elec­tronic sup­ply chain man­age­ment) and ePOS (elec­tronic point of sale). All these de­vel­op­ments have led to hy­per-scal­a­bil­ity, high-avail­abil­ity, re­li­a­bil­ity and high-con­cur­rency to the en­tire process of de­liv­er­ing civil sup­plies in the state, along with low la­tency.

ePDS: a mis­sion crit­i­cal ap­pli­ca­tion

The Post­gres data­base, which cur­rently holds nearly 38 mil­lion records, was de­ployed right at the be­gin­ning of the project. This open source data­base helped to en­able Aad­haar-based bio­met­ric au­then­ti­ca­tion in real-time, and of­fer cash­less as well as card­less trans­ac­tions.

“The ap­pli­ca­tion that we have de­vel­oped with the sup­port of NIC and En­ter­priseDB for dis­tribut­ing ra­tions in the re­gion is very mis­sion crit­i­cal and citizen-cen­tric,” Ra­jasekhar told Open Source For You.

The lat­est Post­gres re­lease (ver­sion 9.6), which has been pow­er­ing the dis­tri­bu­tion sys­tem, can be scaled to more than 1.5 mil­lion trans­ac­tions per sec­ond (TPS). How­ever, the team is work­ing on par­al­lel data pro­cess­ing fea­tures to achieve a higher TPS.

The early chal­lenges

Like with any other large scale project, the de­vel­op­ment of a huge ePDS also faced some chal­lenges ini­tially. As Ashish Nau­riyal, a lead so­lu­tions con­sul­tant at En­ter­priseDB, ex­plained in a con­ver­sa­tion with Open Source

For You, the ma­jor chal­lenges were de­sign­ing, de­ploy­ing and main­tain­ing Post­gres for a mis­sion crit­i­cal, statewide ePDS. “Some of the spe­cific chal­lenges faced were ap­pro­pri­ate Post­gres ver­sions and project-rel­e­vant fea­tures, data­base and server-level tun­ing, ex­pert guid­ance, im­ple­men­ta­tion of best prac­tices, tools re­quired to achieve de­sired data­base-level tasks such as di­ag­no­sis of per­for­mance is­sues, mon­i­tor­ing, high-avail­abil­ity, main­te­nance op­er­a­tions, and on­go­ing data­base server and ap­pli­ca­tion level changes,” re­calls Nau­riyal.

The en­gi­neers lever­aged their ex­ten­sive ex­pe­ri­ence in the data­base man­age­ment space to over­come these chal­lenges. The en­gi­neer­ing team pro­vided sup­port for dis­as­ter re­cov­ery, se­cu­rity con­trols, per­for­mance tun­ing, and back­ups and up­grades in ad­di­tion to de­sign­ing and im­ple­ment­ing sys­tems. Also, the NIC team was al­lowed to change the man­age­ment for the data­base with min­i­mal or even no pro­duc­tion data­base ser­vice dis­rup­tion.

For the team at NIC, the ma­jor ob­sta­cle was the ex­e­cu­tion of tasks like bio­met­ric au­then­ti­ca­tion, the Aad­haar-en­abled pay­ment in­ter­face and hy­per-scal­ing. “With the ap­pro­pri­ate ar­chi­tec­ture and op­ti­mi­sa­tion through the ef­forts of En­ter­priseDB, the so­lu­tion worked out flaw­lessly,” says Ra­jasekhar.

How­ever, as the state gov­ern­ment was com­mit­ted to im­prov­ing trans­parency and ef­fi­ciency, no ma­jor ad­min­is­tra­tive chal­lenges were faced in opt­ing for the ad­vanced ePDS model.

Linux as a soft­ware stack

The NIC en­gi­neers used Linux as the soft­ware stack for the Post­gres de­ploy­ment. This mix is con­sid­ered to be the real ic­ing on the cake. “Post­gres and Linux orig­i­nate from a like-minded

com­mu­nity. This com­mu­nity essence makes Linux an op­ti­mal plat­form for run­ning Post­gres,” says Nau­riyal.

The pres­ence of Linux en­ables ex­perts like Nau­riyal to con­trol the data­base with ease. It also pro­vides the flex­i­bil­ity to choose from mul­ti­ple file sys­tems to en­able sat­is­fac­tory re­sults.

“Con­trol and vis­i­bil­ity of a plat­form, along with the flex­i­bil­ity to choose from mul­ti­ple file sys­tems, help data­base ad­min­is­tra­tors to man­age a Post­gres data­base that is per­form­ing well,” says Ra­makr­ishna Malem­pati, se­nior data­base con­sul­tant, En­ter­priseDB.

Other open source de­ploy­ments

Apart from Post­greSQL, the NIC team also de­ployed Cassandra and Mon­goDB as the data­base so­lu­tions to make its ePDS, eSCM and ePOS suc­cess­ful in the state. The use of all these open source so­lu­tions has re­duced the CAPEX and OPEX of this project to zero.

The suc­cess of AP’s ePDS model has led to other states plan­ning for its repli­ca­tion, with Ra­jasthan and Ma­ha­rash­tra be­ing the first two states to do so.

N. Chan­drababu Naidu, chief min­is­ter, Andhra Pradesh

The ePDS project: Main­tain­ing hu­mon­gous pub­lic dis­tri­bu­tion records in a dig­i­tal form

Newspapers in English

Newspapers from India

© PressReader. All rights reserved.