System requirements

Functional:

1. Multiple device can be added for given user account
2. Activity that can be tracked like: eat, sleep,walk,run, drive, weight
3. Sleep activity data will be collected from the device with sensor, and give detailed report about user's sleep cycle.
4. Running and walk can be tracked by enabling device with acceleratormeter sensors.
5. If user device goes offline, device should keep these data in local cache and sync once decide comes online.
6. If user not wearing a device and not able to track for some duration he should be manually able to add progress for different activity.
7. The application should have a stop timmer for any activity like gym, Swim
8. application can deteact drive timing by using accelerator and GPS tracker.
9. User can set a goal, and system can send remainder about completing the goal
10. System can send motivational cards to complete goal. based on 90% done and 90% pending goal status.
11. user should be able to set and track different level of goals: day, week, month, yearly
12. user should be able to get analysed graph for the activity progress.
13. User can see monthly completed goal. weekly heart beat status graph.
14. Weekly running/swim/walk step graph.
15. User cloud account integration
16. use profiling for different devices.
17. Uses machine learning algorithm to detect and give observation and users activity predections.
18. add remainder for health related days and inspiration to set goals for the same
19. User can follow other user profile to motivate and see each other goals and status.
20. User can ask other user to follow a perticular task/goal
21. if device is not able to track some activity, its status should be marked as not tracked.

Non-Functional:

1. System should always be available
2. The device should track details offline and upload details in bulk once device is online
3. User activity data should be kept in low cost storage like s3 and analysed and refined data can be kept in fast access memory to quickly render user dashboard
4. Deploy services on cloud infra for multiple availability zones for better availability
5. insure encryption for data privacy protection to abide by multiple health agencies guidelines like GDPR(General data protection regulations).
6. Use redis cache for keeping recently tracked activity for fast loading of the profile
7. Use node.js for backend services due to its vast support of concurrent connections, its better suited for real time tracking of user activity via concurrent socket connections.
8. User tensor flow and pyTorch for predective analysis on user historical data
9. Use block chain tech to keep user data integrated and secured using (ZKP) zero knowledge proof concept to validate and share the evaluation result without disclosing the actual user collected data.
10. Sync same matrix from 2 device and give priority to device
11. handle offline updates, when device is offline.

Capacity estimation

Assume total active user: 1 Million

Assume increment of user per year by : 10% i.e 10000 and decrement of 3% per year.

total max device of a user: 2

Device: apple watch 1

Activity log to be sent from device

time: int64

activity: string

start time: long

end time: long

tracking type: String

min heart rate: int

max heart rate: int

This payload size: 2 KB

update sent to server from device:

1. after finishing activity via socket connection to avoid overhead of aggregated uploads
2. approx. size: 5KB per 10 min

Total network call to service: 5 M / 10 Min

Network bandwidth uses range: 2KB - 5KB

1. User profile data can be kept in Postgres

Size: 1 KB size per user

user name: String

user id: String

DOB: String

profile pic url

1. User device data to be kept in postgres

userId

deviceId

device details

size: 0.5KB per user

read request network calls = 5 request per 10 min for active user * 5 Million

• Store active user data for last 1 year in timeseries db like influx DB.

Influx DB storage estimations:

10 Min = 5KB

60 Min = (5*60)/10

1 Yr= (5*6)* 60 * 24 * 365

• Store user analysed and compressed data on s3 for historical data processing and physical health degradation monitoring via LLM

Use AWS glacier or Azure storage to keep least frequently used data.

App uses and performance during peak seasons like new year resolution, world health day etc

API design

Define what APIs are expected from the system...

Database design

Defining the system data model early on will clarify how data will flow among different components of the system. Also you could draw an ER diagram using the diagramming tool to enhance your design...

High-level design

You should identify enough components that are needed to solve the actual problem from end to end. Also remember to draw a block diagram using the diagramming tool to augment your design. If you are unfamiliar with the tool, you can simply describe your design to the chat bot and ask it to generate a starter diagram for you to modify...

Request flows

Explain how the request flows from end to end in your high level design. Also you could draw a sequence diagram using the diagramming tool to enhance your explanation...

Detailed component design

Dig deeper into 2-3 components and explain in detail how they work. For example, how well does each component scale? Any relevant algorithm or data structure you like to use for a component? Also you could draw a diagram using the diagramming tool to enhance your design...

Trade offs/Tech choices

Explain any trade offs you have made and why you made certain tech choices...

Failure scenarios/bottlenecks

Try to discuss as many failure scenarios/bottlenecks as possible.

Future improvements

What are some future improvements you would make? How would you mitigate the failure scenario(s) you described above?