Overview

DataObjects.Net v4.0 is a comprehensive business logic layer development framework providing it with:

• Storage independent data access layer - probably the most independent one. It supports not just SQL databases, but generally any storage that can be classified as indexing engine.
• Object-relational mapping framework providing completely transparent persistence - we've gone even further than with v3.9 here. We support arbitrary Keys, persistent Entities, Structures and EntitySets.
• Common services you'll need in almost any case, such as:
  • Validation
  • Atomicity
  • Transactional services
  • Full-text search
  • Caching
• A set of our own storage providers allowing you to get rid of SQL databases at all. Why? Because we think SQL is dying. All of them are indexing engines supporting RSE - they index the data and allow to query them. They're transactional. So they offer nearly the same features as any SQL database does, but on the other hand, they eliminate many levels of abstraction (e.g. SQL itself, and thus - its parser, query optimizer and so on) and provide open indexing API allowing to deal with them almost directly. In particular, we're working on:
  • Memory storage - it stores the data in RAM. Quite fast. Perfect for client-side caching and unit tests.
  • File system provider - in fact, a regular database. Not available yet, but this is one of features we're going to deliver ASAP.
  • Distributed storage provider - upcoming distributed storage implementation.

DataObjects.Net v4.0 is successor of DataObjects.Net v3.9, but on the other hand, it is a complete rewrite of it:

• It is designed basing on new concepts
• It uses benefits of the most recent version of .NET 3.5 and C# 3.0
• It is written in accordance with much stronger coding and designing standards.

What makes DataObjects.Net different?

It really differs from any existing ORM tool at this moment. The key differences are:

• It has built-in database engine (Indexing engine), which, in particular, allows to:
  • Use it without any external database engine at all
  • Support n-tier architecture much better - any node in node chain can query its local storage in the same way without any performance drawbacks. So your middle-tier server can execute queries over its own storage, or your DTOs can find the others using queries.
• It is designed to:
  • Support not just SQL databases, but generally any database that is able to expose its indexes
  • Support SQL databases much better - SqlDom brings a new level of abstraction allowing to add support for almost any new SQL database in days
  • Work with distributed storage. In particular, its query compilation layer is designed in a way allowing to compile a part of query independently basing on the current configuration of the distributed storage (data server cloud).
• It provides new implementation of such formerly well-accepted levels of abstraction, as:
  • Atomicity - now we use PostSharp to gather method invocation data and provide undo-redo. Atomicity framework replaces formerly existing nested transactions.
  • Transactions - all is nearly the same here, but now - much more simple and safer
  • We also brought a new Validation framework, that DataObjects.Net v3.9 was missing so much.

What benefits does it bring?

You get:

• IMDB (in-memory database) - so it must help to make your smart client applications (e.g. WPF applications) better.
  • Running unit tests on IMDB is one of the most awaited features we got for DataObjects.Net v3.9.
• Embedded database engine - it's a part of such applications, as Firefox, Thunderbird, ICQ, Skype
• ORM tool - it must help you to increase the speed of development at least twice.

These components are essential in building an application interacting with a database. So we provide all-in-one package.

In addition, you get very large library (check out a list of namespaces) of additional classes solving such problems, as (just some examples of many):

• Generic arithmetics ( + , - , / , * , etc. for arbitrary T type - quite useful if you write e.g. generic sequence generator)
• Xtensive.Core.Serialization - much more advanced serialization
• TopDeque class - very useful structure that is missing in .NET - a hybrid of Dictionary and Queue
• Xtensive.Core.Caching - a set of frequently used caches, including LruCache, MfLruCache, WeakCache and WeakestCache
• Xtensive.Core.Diagnostics - log4net-based logging and Measurement class
• Xtensive.Core.Aspects - a set of generic PostSharpaspects, such as ImplementAutoPropertyReplacementAspect class or ImplementConstructorAspect

All features listed above are parts of Core project. Check out other projects to get the imagination of what's left.

See also

• Concepts - the concepts behind DataObjects.Net v4.0
• SQL is dying - an article describing the one of the key concepts in DataObjects.Net v4.0 design