Beautiful Data

Updating Data

As users interact with websites, their actions constantly result in database updates. The first challenge we examine is how to support this massive stream of updates while providing good performance and consistency for each update.

The Challenge

Imagine that we want to build a social networking site. Each user in our system will have a profile record, listing the user's name, hobbies, and so on. A user "Alice" might have friends all over the world who want to view her profile, and read requests must be served with stringent low-latency requirements. For this, we must ensure that Alice's profile record (and similarly, everyone else's) is globally replicated so those friends can access a local copy of the profile. Now say that one feature of our social network is that users can update their status by specifying free text. For example, Alice might change her status to "Busy on the phone," and then later change it to "Off the phone, anybody wanna chat?" When Alice changes her status, we write it into her profile record so that her friends can see it. The profile table might look like Table 4-1. Notice that to support evolving web applications, we must allow for a flexible schema and sparse data; not every record will have a value for every field, and adding new fields must be cheap.

Table 4-1. User profile table

Username	FullName	Location	Status	IM	BlogID	Photo	…
Alice	Alice Smith	Sunnyvale, CA	Off the phone, anybody wanna chat?	Alice345			…
Bob	Bob Jones	Singapore	Eating dinner		3411	me.jpg ...