Managing Timezone & Dst Issue For Javascript Application

I am trying to create a scheduling application. The front end\UI is developed using JavaScript. The back end is a ASP.NET Web Api application which uses MSSQL server as the databas

Solution 1:

Warning - Scheduling properly is hard. There's a lot more to consider. Please read this and this. Most of your questions are addressed there (though from the perspective of other languages, the challenges are the same).

You might also take a look at Quartz.net, which is sufficient for many scenarios.

To answer your specific questions:

1. How do I store time information (10 AM in this case) on SQL server?

For the recurrence rule, store the local time of the event. SQL Server has a time type, which works well for storing the time of day. You will need other fields for tracking the time zone, the start date, days of the week, and other pattern information.

For the specific instance that is scheduled to run, you calculate the UTC datetime based on all the information in the recurrence rule. At minimum, you schedule the next occurrence, and recalculate after each run. In some cases, you may decide to project the next N occurrences, depending on what you need to show to the users. (You could also use a datetimeoffset for this purpose. See datetime vs datetimeoffset.)

2. Should I get the time using JavaScript on client machine and then convert the same to UTC?
3. Should I get the time using JavaScript and also save the user time zone information?

To answer both questions: For scheduling, you should not discard the original input, which will be in the local time zone of the event being scheduled. That may or may not match the time zone of the user. You will need to ask the user to select the time zone of the event.

4. What happens when DST related changes take effect?

That's up to you. You will need to test this thoroughly. In general, there is a period of local time that is skipped, and a period of local time that is repeated.

• When it is skipped, you have to decide when to run the event. Options include: 1) before the skipped time, 2) after the skipped time, and 3) not at all. In most cases, the preferred option is to run after the skipped time, by advancing the local time by the DST bias (usually 1 hour). For example, a daily event scheduled to run at 2:30 every day in Pacific time would run at 3:30 on the day of the spring-forward transition.

• When it is repeated, you have to decide when to run the event. Options include: 1) at the first occurrence, 2) at the second occurrence, and 3) at both occurrences. In most cases, the preferred option is to run at the first occurrence only. For example, a daily event scheduled to run at 1:30 every day in Pacific time would run at 1:30 PDT, and not at 1:30 PST.

  • Exceptions to this include dealing with businesses that are open late into the evening and choose to stay open for the repeated hour. For example, a bar, restaurant, or movie theater. It is highly dependent on the specific use case and the choices made by the specific business.

5. Will library like moment.js will help in this scenario?

Not from a scheduling perspective, no. It can help with parsing, formatting, and validating input though. You might also use moment-timezone to help with selecting the event's time zone. If you were running this with node.js on the back end, then perhaps there would be more benefit.

---

The biggest challenge is actually one you have not talked about, which is maintaining the time zone data on your server. In your C# code, I recommend using Noda Time for this, instead of TimeZoneInfo. You can then update the tzdb data yourself as needed. You also need to think about the workflow of rescheduling the UTC instants of each occurrence, in the case that a time zone has changed its offset or daylight saving time dates.