保持内存中的更改与后端同步的好方法是什么

Question

What is a good way to keep changes to a resource in sync with both in-memory and persistent storage in a multi-user situation? Here's two ideas, but they're both clumsy

Candidate A

Here, a copy of r must be retrieved so that in-memory state and persistent state is updated at the same time on put .

Resource r = resources.getDeepCopyOfResourceById("foo");
r.removeValue("bar");
resources.put(r);

Candidate B

Here, update operations are located on the service level. This appear to me as an object-oriented anti-pattern.

Resource r = resources.getResourceById("foo");
resources.removeValue("foo", r);

Is there a better way to design this?

Answer 1

What is a good way to keep changes to a resource in sync with both in-memory and persistent storage in a multi-user situation?

Generally speaking you have two problems here: Caching and concurrency.

For caching you need to decide what strategy to use. There are plenty of options. For example:

• read through / write through
• write around
• write behind
• cache aside

When you've decided the caching pattern the next step is to make the cache and the persistent data access thread safe. This is usually achieved with transactions and locks. The exact implementation depends on which libraries you use.