Synopse

In a discussion with Henrick Hellström, in Embarcadero forums, I wrote some high-level information about mORMot.

It was clear to me that our little mORMot is now far away from a simple Client-Server solution.

The Henrick point was that with Real Thin Client (RTC), you are able to write any Client-Server solution, even a RESTful / JSON based one.

He is of course right, but it made clear to me all the work done in mORMot since its beginning.
From a Client-Server ORM, it is now a complete SOA framework, ready to serve Domain-Driven-Design solutions.

Did you hear from the great Smart project?

It is an IDE and some source runtime able to develop and compile an Object-Pascal project into a HTML 5 / CSS 3 / JavaScript embedded application.
It does target AJAX Mobile application creation (i.e. Android and iPhone/iPad apps running Web-Kit).
You'll get an unique .html file containing the whole client-side application: it won't need any server side implementation. Using a third-party tool like PhoneGap, you'd be able to supply your customers with true native applications, running without any network, and accessing the full power of any modern Smart Phone.

Smart is a great candidate for implementing rich client-side AJAX applications, to work with our client-server mORMot framework.

In order to interface Smart code with mORMot, we started implementing some low-level code to work with our RESTful authentication scheme.

So we'll need to implement some Smart dedicated Open Source code implementing crc32 and SHA-256 hashing.

One nice feature of the TDynArray wrapper and its associated methods, as defined and used in our framework, is the ability to serialize any dynamic array as JSON content.

By default, only "standard" dynamic arrays (like TIntegerDynArray) are serialized as true JSON array: other not known kind of arrays are serialized as binary content, within a Base64 encoding.

This is a very efficient solution for a pure Delphi application, since it will be fast and always works, but won't be easy to deal with from an AJAX client.

Applications can now supply a custom JSON serialization for any other dynamic array, via the TTextWriter.RegisterCustomJSONSerializer() class method.
Two callbacks are to be supplied for a dynamic array type information, in order to handle proper serialization and un-serialization of the JSON array.

As stated by this previous article, the default answer format is a valid JSON object.

In some cases, it may be useful to have a service operation (i.e. an interface method) returning any content, e.g. some plain TEXT, HTML or binary data (like a picture).

You will find out in SQLite3Commons.pas all classes implementing this interface communication.

There are two levels of implementation:
- A services catalog, available in TSQLRest.Services property, declared as TServiceContainer (with two inherited versions, for each side);
- A service factory for each interface, declared as TServiceFactory (also with two inherited versions, for each side).

In fact, TServiceFactory.Create constructor will retrieve all needed RTTI information of the given interface, i.e. GUID, name and all methods (with their arguments). It will compute the low-level stack memory layout needed at execution. And the corresponding "contract" will be computed, to validate that both client and server expect the exact same interface.

On the server side, TServiceFactoryServer.ExecuteMethod method (and then a nested TServiceMethod.InternalExecute call) is used to prepare a valid call to the implementation class code from a remote JSON request.

On the client side, a TInterfacedObjectFake class will be created, and will emulate a regular Delphi interface call using some on-the-fly asm code generated in the TServiceFactoryClient.Create constructor.

In order to have an operating service, you'll need to implement a Delphi class which matches the expected interface.

The Client-Server services via methods implementation (our DataSnap-like feature) gives full access to the lowest-level of the mORMot's core, so it has some advantages:
- It can be tuned to fit any purpose (such as retrieving or returning some HTML or binary data, or modifying the HTTP headers on the fly);
- It is integrated into the RESTful URI model, so it can be related to any table/class of our ORM framework (like DataAsHex service above), or it can handle any remote query (e.g. any AJAX or SOAP requests);
- It has a very low performance overhead, so can be used to reduce server workload for some common tasks.

But this implementation pattern has some drawbacks:
- Most content marshaling is to be done by hand, so may introduce implementation issues;
- Client and server side code does not have the same implementation pattern, so you will have to code explicitly data marshaling twice, for both client and server;
- The services do not have any hierarchy, and are listed as a plain list, which is not very convenient;
- It is difficult to synchronize several service calls within a single context, e.g. when a workflow is to be handled during the application process (you have to code some kind of state machine on both sides);
- Security is handled globally for the user, or should be checked by hand in the implementation method (using the aParams.Context values).

You can get rid of those limitations with the interface-based service implementation of mORMot. For a detailed introduction and best practice guide to SOA, you can consult this "classic" article.

According to this document, all expected SOA features are now available in the current implementation of the mORMot framework (including service catalog aka "broker").

Here is the definition of "cache", as stated by Wikipedia:

In computer engineering, a cache is a component that transparently stores data so that future requests for that data can be served faster. The data that is stored within a cache might be values that have been computed earlier or duplicates of original values that are stored elsewhere. If requested data is contained in the cache (cache hit), this request can be served by simply reading the cache, which is comparatively faster. Otherwise (cache miss), the data has to be recomputed or fetched from its original storage location, which is comparatively slower. Hence, the greater the number of requests that can be served from the cache, the faster the overall system performance becomes.

To be cost efficient and to enable an efficient use of data, caches are relatively small. Nevertheless, caches have proven themselves in many areas of computing because access patterns in typical computer applications have locality of reference. References exhibit temporal locality if data is requested again that has been recently requested already. References exhibit spatial locality if data is requested that is physically stored close to data that has been requested already.

In our ORM framework, since performance was one goal since the beginning, cache has been implemented at four levels:

• Statement cache for reuse of SQL prepared statements, and bound parameters on the fly - note that this cache is available not only for the SQlite3 database engine, but also for any external engine; 
• Global JSON result cache at the database level, which is flushed globally on any INSERT / UPDATE; 
• Tuned record cache at the CRUD/RESTful level for specified tables or records on the server side; 
• Tuned record cache at the CRUD/RESTful level for specified tables or records on the client side.

The prototype of these methods has been modified one more time, to supply an unique parameter:
This is a CODE BREAK change and you shall refresh ALL your server-side code to match the new signature.

This unique parameter will let the signature remain untouched in your code implementation, even if the framework evolves (like adding a new parameter).