[DRAFT] A note on the Open Network for Digital Commerce (ONDC)

1

Why the ONDC Must be Open Source

The Open Network for Digital Commerce (ONDC) in India was launched with great fanfare by the government and its architects led by Mr Nandan Nilekani on 29th April. The stated goal of the network is to “democratise fast growing e-commerce sector, help small retailers and reduce dominance of online retail giants” by creating a level playing field like the Unified Payments Interface (sic).

This is yet another national governance project that comes on the heels of massive successful implementations of the Universal ID (Aadhaar), the Goods and Service Tax Network (GSTN) and the Unified Payments Interface (UPI). These projects have arguably unlocked the possibility to build large scale platforms (telecom and fintech) and also made payments cheap and affordable to everyone. There have been huge benefits to the public in terms of access and cost and have helped India leapfrog to one of the most technologically advanced administrations globally, making it one of the rare instances where government innovation has raced ahead of both the private sector and academia.

On the flipside, these platforms that affect the lives of a billion people, leave a lot to be desired when it comes to adhering to the essential tenets of openness and democracy. Most of the platforms are controlled centrally with a layer of tightly controlled entities and large players. This has led to the rise of a new set of incumbents who have been at the forefront of harnessing these platforms to build large telecom and fintech platforms. Was there an alternate approach to building out these platforms?

Large decentralised networks

We can look at the Internet as an example of how to build a large decentralised network. There are broadly two types of computers on the Internet, those that are reliably ON and have a near permanent (leased) presence on the network (servers) and those that connect time-to-time based on need (clients). Servers form the backbone of the internet and are identified by a number and optionally a name. These numbers and domain names are given out by an autonomous agency via sub-registrars for a small fee. Based on this simple numbering and naming system and a protocol for connecting and exchanging information and a unified client (browser), the internet has become the largest and the most impactful platform ever created by humanity.

Features that make the internet powerful and inclusive:

  1. The function of the central node is extremely simple (give out names and numbers)
  2. The registration is distributed through a federated network of registrars
  3. It is extremely cheap and easy to set up a server on the internet.
  4. Setting up a server on the internet and the core infrastructure of the internet can be done completely by free and open source software (FOSS)
  5. There is a free unified way of consuming information and applications on the net (web browser)

Each of these features are critical to the runaway success of the internet. The system is massively decentralised and it is extremely easy for anyone to lease a number and use FOSS based servers to start running a service on the internet.

The case for decentralisation

As India prepares its next phase of national platforms, it would be important to reflect on the architectures promoted by these projects. Can we envisage these platforms as parts of society (like the internet) rather than as parts of government? How easy or hard is it to become an ID provider, a payment service provider? Just as I have an option to run my own website or email account, do I have the option of hosting my own wallet?

While some of these questions may sound naive in the face of internet scale security threats and frauds, the internet has proved resilient to all kinds of large scale attacks right from spamming to service attacks. At the same time it has opened the floodgates for anyone with talent to set up their own internet service without being beholden to any bureaucracy. We believe that India’s national infrastructure must make it easy for anyone with talent and minimum resources to be able to participate by using the platforms to create a new generation of services and platforms on top of it.

As we look at the ONDC, a few key questions spring up:

  1. How easy is it to register and run a marketplace?
  2. How easy is it to register and run my own shop?
  3. Do we have open source implementations of marketplaces and merchant tools so that people can easily participate without signing up with a 3rd party?

While the Beckn Protocol (on which we assume ONDC is based) aims to be an internet scale protocol, without free and open source implementations, it has the potential of turning into a massively centralised system that would end up benefiting a few incumbents. Hence our strong recommendation would be to open source every implementation of this protocol and co-create along with the community so that we can build a truly decentralised national platform that will be robust and resilient and truly democratic.

Note: These are just my personal views, not officially of FOSS United (maybe we need a mechanism for that)

1 Like
2

Gotta say, government projects aren’t to be judged like private projects. Government is not a private entity. In Malayalam they say it is not anyone’s ancestral inheritance.

3

How do you propose government projects be judged?

In a democracy, the people are sovereign, so in some equation, it is our responsibility to make our government accountable. Either that or I am not getting your point :smile:

4

Is ONDC a public project in character?

I have a problem in calling ONDC a national governance project / associating with government or for that matter “decentralised national” platform.

There has been constant media misrepresentation / misinformation and the duality / careful articulation of ambiguity between public / private as per convenience of those para-state actors who govern the digitization agenda has been the root of many problems in digitalization (Aadhaar, GSTN, NPCI, NDHM - insert your favourite ‘volunteer-ware’).

On the one hand - Department for Promotion of Industry and Internal Trade (DPIIT) initially setup committee / ‘advisory council’ on the issue, it is important to note “Open Network for Digital Commerce” - ONDC is a private sector led initiative. In addition to ONDC being registered as a private non-profit, note the careful omission of the word government in response to a parliamentary question by the Minister. Just as the Prime Minister’s photo on frontpage advertisements in newspapers for a fintech / telecom, doesn’t make those entities “government”, a minister reading a statement about an entity in parliament that is privately owned with some blessing by his department - doesn’t make the entity a government one / define its public character. Those of us watching the space know the interests behind the “technical protocols” and as with the history of open-washing, the new trick (or I must say, the old trick, that many of us never realized) is public-washing - where the entity will pose as a quasi government owned entity - when it comes to matters of building public opinion / showcasing the legitimacy using authority / power that is derived using bureaucrats / government arms - but act entirely as a private entity when it comes to matters of transparency, accountability, social justice.

A non-profit monopoly with power - still hijacks / centralizes power (even if not data) - and such power centralization in private institutions is an early indicator on many things that will go wrong. An example of public / private chameleonism from the past decade was IL&FS which led to some ~ 50,000 crore scam. A selective government blessing to one such private sector led initiative is fundamentally anti-comeptitive and picks winners. This has already happened in the payments landscape and the cost of it is being borne by the people - while the benefits of such decision are being enjoyed by those para-state digitization drivers who got the government to back them.

A public good by defnition ignores social harms it causes.

Decentralisation and FOSS friendliness sufficient?

India Stack must be free, open. - TL;DR - The critical difference even if IndiaStack were F/OSS - is that the barriers to entry are far high - usually gatekept by the same folks who build. Decentralisation again is not sufficient - if there is a mandatory interconnect reguation - under the garb of interoperability / some regulation. Because once the data flows - its just one step closer to market capture. This is sort of right to disconnect to prevent - “One Nation One Market One Seller” if I may.

How hard it is to become a payment service provider?

In 2018, a draft bill was made by Departmet of Economic Affairs to make payments as a seperate sector outside of RBI, but the powers that govern, made sure the bill never got tabled in parliament. 3 years and 2 consultations later ( Consultation on Authorisation of Retail Payment Systems, Draft Framework for authorisation of a pan-India New Umbrella Entity (NUE) for Retail Payment Systems ) , RBI is slowly killing that possibility. While its possible to get a PPI license on paper, the mandatory interconnect with UPI would mean - there cannot be differentiated offering and mandatory interoperability using the preferred platform sets a kind of monoculture.

In its current protocol design - Beckn enables Friend of States™ (ONDC) to centralize “demand data” while leaving the fulfillment data with the private sector (buyer / seller apps). This will inherently create an arbitrarge as those with demand will be able to deliver efficiently at lower cost / build infra using demand data(or that’s the promise that is being sold)

There are limitations with a monopoly private non-profit having this arbitrarge and any mandatory F/OSS solutions will not change this power differential.

What can be done though?

Having said these, I am not reducing the need for F/OSS tooling, but they have to be complemented with “Watchdog tooling” and policy feedback loops (which eventually gives public a stake in platform governance and not limits to investors of private non-profit ONDC) that prevents / minimises social harms as it is fair to assume those with economic interests at stake will justify social harms with public good narrative building - as the case has been in the last decade.

Addendum :- An twitter thread on “Aadhaar doublespeak” / Aadhaar speak about ONDC is at https://twitter.com/logic/status/1529365728649891840

3 Likes
5

By using the Constitution of India.

1 Like
6

What can done - shall be done. GitHub - DigitalIndiaArchiver/NammaYatriStats: NammaYatri Stats Archiver archives statistics of NammaYatri - the mobility implementation of Beckn / ONDC. Someday - when this becomes monopoly - we will have time-series view of market capture - even if the proponents decide to take these opendata down.

My political observation of these opendata (particularly - following payments open data - which NPCI / even PhonePe publishes) coming out from modern platforms is - these are carefully vetted datasets by its owners that doesn’t highlight any negative aspects of the platforms. For instance - the NammaYatri opendata - will never tell someone about say - women safety.

Counter intuitively, As the saying goes - If you torture data enough, it will confess - there is probably something that this data set confesses that is up for extraction. For someone with local knowledge / context - its even easier (If you are regular user from those low service NammaYatri areas) to make the data confess / look at right areas to extract data that can signal harms / tell areas for improvement.