Lightspeed 2nd level caching - clarification request

Hi, I'm using LIghtSpeed v5. I've just read the documentation on the 2nd level caching. There are a few things I didn't understand:

1. Which of the following is true: (a) Enabling 2nd level caching via code (e.g. context.Cache = new CacheBroker(new DefaultCache());) or via the config file (setting '<'lightSpeedContexts'>'...) is enough to de facto activate it for all entities (as opposed to me having to turn the caching on for each individual entity). or (b) Setting the "Cached" property to "true" for an entity in the designer will activate the 2nd level caching for that particular entity, even if I didn't set "context.Cache" property in the code or in the config.

2. If (a) is above the case (i.e. 2nd level caching will be enabled for all entities if "context.Cache" property is set), what is the default expiry period for the entities in which custom cache expiry period is not specified? Also, if (a) is the case, what will turning caching on for the individual entities (i.e. setting the "Cached" property to "true" in the designer) actually do?

3. How can I know during the runtime, if/when and for what entities the 2nd level cache was activated? Using LightSpeed logging maybe, or with the use of the debugger?

4. Is there any way to set caching parameters on entities (e.g. whether they are being cached and their expiry period) programmatically or via an external config file (NOT in the model designer)?

5. How does the cache expiry mechanism work: let's say I have 2nd level caching enabled for the table "Units", and its expiry period set to 1 minute. Now, I've selected a record using Units.FirstOrDefault(x => x.Id == 5) in the context of UnitOfWork1. After a while I make a change to this unit and call UnitOfWork1.SaveChanges(). My question is: will the 2nd level cache be updated with the change that I've made, so that if I call Units.FirstOrDefault(x => x.Id == 5) on UnitOfWork2 within the span of 1 minute, the resulting Unit entity will see the changes I've just saved, even though it was taken from the cache (and not through an SQL query)?

6. Concerning the entity cache expiry period parameter: when does the countdown to zero begin (zero being the point in time when the entity is deleted from the 2nd level cache)? Right after the entity is first queried from the DB and loaded into memory or, say, only after the unit of work in the context of which it was queried is disposed of?

Apologies for this many questions. Thanks in advance!

Hi,

Sorry about the delay in replying.

Before diving into your specific questions, it may help to know that the Caching provider for LightSpeed is a pluggable model. This means you can write your own caching provider very easily with your own settings depending on your needs. By default, the included caching provider uses the .NET HTTPCache (note: you don't actually need to be using ASP.NET or similar to use that cache, it's very good but tucked in the web related namespaces).

That may help give some guidance on how our caching works, but to answer your specific questions:

1. Second level caching applies only to entities attributed with the Cached attribute AND you have set a 2nd level cache provider. Note: The first level cache, in memory with the UnitOfWork, applies globally all the time. Once and entity is attached to the UnitOfWork it will be re-used as needed.

2. All of the defaults associated with the HttpRuntime.Cache apply. This means that it will never expire, however when memory goes low or there's more valid data to be stored, the ASP.NET cache will remove the item (read more here: http://stackoverflow.com/questions/1423880/default-duration-of-cache-insert-in-asp-net)

3. The cache will be used if it's set up, and if the data is still in there (which we don't control when it expires, see previous questions). Then fetching objects will result in a check if we already have it in the cache and re-use that entity.

4. Not specifically - the designer is just doing code generation, but you'd still want to use the designer. You could, in theory, write your own cache provider very simply and add logic in that to lift config from somewhere however.

5. You would get the up-to-date version. When saving it updates the cached value.

6. The count-down is from the time it is added to the cache. This will be when it's queried and pulled out from the database. Not when the UoW is disposed.

I hope that helps!

John-Daniel Trask

Thanks John-Daniel, it did help a lot!