Managed connector release

Owned by Vineet Singh, created
Jun 24, 2021
3 min read

Add your comments directly to the page. Include links to any relevant research, data, or feedback.

Status

In Progress

Impact

High

Driver

@Vineet Singh 

Approver

@rikin (Unlicensed)

Contributors

@Stuti Verma (Unlicensed) @Razak K M @Mayank (Unlicensed) 

Informed

@saureen (Unlicensed)

Due date

Jun 24, 2021

Outcome

Background

There are currently two ways farmstack can be used:

  1. Someone builds it using github code, being completely open source we have no issues

  2. Digital Green offers FarmStack installer files that can be used to setup FarmStack by respective users

However, there might be users who do not want to run the installer but use FarmStack connector for usage policies and create a connector which is not on their infrastructure. This is managed connector which gives users an ability to fire up connector from the browser, integrate with their applications (source) and specify destinations with associated end applications.

The decision is to come up with a mock use case which helps users experience this on farmstack/ digital green website. The objective is not to create an elaborate product but a clean and simple user experience that shows ease of using the connectors.

Decision criteria

Decide on the basis of:

  • ease of effort for non-farmstack activities

  • ease of explanation for audience

  • usage policy configuration

Options considered

 

Option 1: Managed flow triggering an action as per the usage policy

Option 2: Mock data integration as per the usage policy

 

Option 1: Managed flow triggering an action as per the usage policy

Option 2: Mock data integration as per the usage policy

Description

Visitors on farmstack website enter details about themselves including email and phone number with crop and pre-defined list of places.

They configure usage policy that their data is available for a) an application that recommends videos on their email and b) an application that recommends videos on their whatsapp/ telegram

The application itself doesn’t store data.

Two files with mock data is generated for each visitor which has sensitive information about the farmers. The files need to be combined together to create analytics like time series and pie charts by a third user. The consumer just wants to join the two files based on the common identifier but does not want to store the sensitive information either.

An application that does joining and filtering is created by the consumer and the connector helps configure the policy that gives access only to this application.

User flow

Visitors come on website and they follow the following step:

  1. CTA: See FarmStack in action

  2. Sign up form

  3. Use case description with continue button to enter three step model

    1. Step 1: enter details about self (phone number, email, select location form drop down)

    2. Step 2: Pair with an example connector

    3. Select policy: a) available only to application which can mail or b) available only to application that can send messages on whatsapp/ telegram

  4. After setup - user gets an email/ message as configured

  5. The user is explained what just happened

Visitors come on webstie and the follow the following step:

  1. CTA: See FarmSTack in action

  2. Sign up form

  3. Presented with two mock files - for convenience shown in google sheets

  4. Sheet 1 is footfalls on the shop:

    1. who: farmer details (phone, name, farmer residence details etc)

    2. where: date, shop id, village

    3. what: crop, quantity, grade

  5. Sheet 2 is video screen and adoption details:

    1. who: farmer details

    2. what: screennig, adoption

  6. User sees the preview of the application and configure which tables can be shown

  7. Tables like time series - number of farmers by days, efficacy by region - number of farmers for grades A,B, C in respective regions and so on

Pros and cons

  1. Brings together content, chat bot kind of stuff with FarmStack

  2. Focus on tangible outcome

  1. Difficult to explain - takes focus off form farmstack

  2. Difficult to explain that the mailing or messaging application that the containerised application calls is not compromised

  3. Not of much value for audience from US, that is, outside India or Ethiopia

  1. Focus on data joining

  2. Usage policy is clean

  3. Data is controlled by demo

  4. All audience visiting our website can understand

  1. Static datasets - no dynamic thing involved

  2. Need to explain what usage policy did

Estimated effort

Medium Large

Medium

Action items

Discuss in our check in and finalize the flow and decide if we need to add something for the suage policy
Do we need mock ups

Outcome