Case study
Budget: ₹100 – ₹400 INR
CASE STUDY
Company deploys a large number of edge-computing enabled assets like advanced lithium ion batteries, chargers,battery automatic vending machines, etc to enable the energy ecosystem of India with green miles. To reduce therisk of assets like lithium ion batteries, location tracking and capability to manipulate such assets is required.
Continuous access to asset data also helps predict it’s life and helps in pre-emptive management of downtime.
Your mission should you choose to accept is to design an architecture which should account for the below mentioned caveats and requirements
1. SECURITY OF IOT-SERVER CONNECTION
To prevent rogue attacks on the IoT network, security of communication between the asset and server is required.
This requires Detecting Unusual Behaviour Of A Particular Socket(MQTT based or TCP/IP) or Whitelisting like measures to be deployed
2. QUOTA CONSTRAINTS OF AN IOT-SERVER CONNECTION
To prevent undesirable events of instances shutting down due to overhaul or unexpected billing events.
3. DATA COLLECTION/TELEMETRY OVER IOT-SERVER CONNECTION (like TCP/IP WEBSOCKET)
As mentioned before to reduce risk, asset health monitoring is required. It helps prevent undesirable events like battery explosion. Since the Management system inside an asset may vary in a lot of ways including protocol layer(UART, CAN,etc) or physical layer or even in the way it communicates with the server, a support for at least Streaming data type and Request Response data type communication needs to be supported by Company backend.
4. ASSET CONTROL
For controlling the asset like switching off power of battery remotely or manipulating its power giving capabilities or changing the charging profile of charger based on temperature and current life of battery that is being plugged into the charger , a reliable asset control system is required which maintains queues of commands to be sent, acknowledges , prioritizes and maintains state of control commands to track retries
5. RESPONSE PARSER
Given a variety of assets with different combinations of management systems and IoT’s deployed it is imperative to have a response parser which provides communication uniqueness, Processes streams of events with joins, aggregations, filters, transformations, etc, using event-time and exactly-once processing for live dashboards.It should also support extremely high throughputs.
6. STORAGE
Creating data pipelines to store data reliably and efficiently enables two types of use cases. Firstly, queries fast enough to power live dashboards used by partners all across the energy ecosystem including financial partners, OEM’s ,etc. Secondly, data pipelines also need to be established to power AI/ML algorithms to learn and predict.
7. VIRTUAL BATTERY
Required for testing purposes Some numbers to help you get the idea of scale.
S.No. Data point Value
1 Messaging frequency 1000000 messages/ minute
2 Number of active hours per day 24
3 Size per message (KB) 0.5
4 Analytics/Message Retention
duration (days)
a Raw data Infinite
b Aggregated data Infinite
5 Number of connected devices
a Current scale (March 2022) 15000
b March 2023 110000
sample payload for integration - 1 000000000000002a0c010600000022dd03001b115af64
c02eb12a20015273b0000000000002810030e020cb50
c4ffab977010000604f
sample payload for integration - 2 000000000000002b0c010600000023dd04001c103810
371033103810361034103610351035103e1031103910
391036fc097701000006a1
sample payload for integration - 3 000000000000005c8e010000017c7e95ad98002dee737
a111ef13400cb01190c00190000000e000700ef0100f00
100500100150400c800004501007164000600b500060
0b600030042d8bc001800190043101f00440000000100
f100009dda00000000010000451
Company deploys a large number of edge-computing enabled assets like advanced lithium ion batteries, chargers,battery automatic vending machines, etc to enable the energy ecosystem of India with green miles. To reduce therisk of assets like lithium ion batteries, location tracking and capability to manipulate such assets is required.
Continuous access to asset data also helps predict it’s life and helps in pre-emptive management of downtime.
Your mission should you choose to accept is to design an architecture which should account for the below mentioned caveats and requirements
1. SECURITY OF IOT-SERVER CONNECTION
To prevent rogue attacks on the IoT network, security of communication between the asset and server is required.
This requires Detecting Unusual Behaviour Of A Particular Socket(MQTT based or TCP/IP) or Whitelisting like measures to be deployed
2. QUOTA CONSTRAINTS OF AN IOT-SERVER CONNECTION
To prevent undesirable events of instances shutting down due to overhaul or unexpected billing events.
3. DATA COLLECTION/TELEMETRY OVER IOT-SERVER CONNECTION (like TCP/IP WEBSOCKET)
As mentioned before to reduce risk, asset health monitoring is required. It helps prevent undesirable events like battery explosion. Since the Management system inside an asset may vary in a lot of ways including protocol layer(UART, CAN,etc) or physical layer or even in the way it communicates with the server, a support for at least Streaming data type and Request Response data type communication needs to be supported by Company backend.
4. ASSET CONTROL
For controlling the asset like switching off power of battery remotely or manipulating its power giving capabilities or changing the charging profile of charger based on temperature and current life of battery that is being plugged into the charger , a reliable asset control system is required which maintains queues of commands to be sent, acknowledges , prioritizes and maintains state of control commands to track retries
5. RESPONSE PARSER
Given a variety of assets with different combinations of management systems and IoT’s deployed it is imperative to have a response parser which provides communication uniqueness, Processes streams of events with joins, aggregations, filters, transformations, etc, using event-time and exactly-once processing for live dashboards.It should also support extremely high throughputs.
6. STORAGE
Creating data pipelines to store data reliably and efficiently enables two types of use cases. Firstly, queries fast enough to power live dashboards used by partners all across the energy ecosystem including financial partners, OEM’s ,etc. Secondly, data pipelines also need to be established to power AI/ML algorithms to learn and predict.
7. VIRTUAL BATTERY
Required for testing purposes Some numbers to help you get the idea of scale.
S.No. Data point Value
1 Messaging frequency 1000000 messages/ minute
2 Number of active hours per day 24
3 Size per message (KB) 0.5
4 Analytics/Message Retention
duration (days)
a Raw data Infinite
b Aggregated data Infinite
5 Number of connected devices
a Current scale (March 2022) 15000
b March 2023 110000
sample payload for integration - 1 000000000000002a0c010600000022dd03001b115af64
c02eb12a20015273b0000000000002810030e020cb50
c4ffab977010000604f
sample payload for integration - 2 000000000000002b0c010600000023dd04001c103810
371033103810361034103610351035103e1031103910
391036fc097701000006a1
sample payload for integration - 3 000000000000005c8e010000017c7e95ad98002dee737
a111ef13400cb01190c00190000000e000700ef0100f00
100500100150400c800004501007164000600b500060
0b600030042d8bc001800190043101f00440000000100
f100009dda00000000010000451
Related categories:
Cloud Computing