Do Cloud Vendors ensure ACID properties while offering Relational Database services?

AWS RDS looks to be based off of mysql 5.1, and SQL Azure is based off of SQL Server. Both of those platforms are ACID compliant, seems a fair bet that RDS/SQL Azure are as well.

These offerings are scalable not by virtue of architecture (like most key/value stores); but simply because they can offer big honking servers if you pay for them, and I guess they both have a well-tuned installation and experienced staff for keeping them running.

As Tony said, they're based on traditional SQL servers (MySQL andMSSQLServer, respectively), make no mistake, this is real software and can really deliver when properly tuned on appropriate hardware.

Actually i read a good paper which talks about my question, i wanted to post here for further users. From DAVID CHAPPELL OCTOBER 2008, INTRODUCING THE AZURE SERVICES PLATFORM AN EARLY LOOK AT WINDOWS AZURE, .NET SERVICES, SQL SERVICES, AND LIVE SERVICES

SQL Data Services, formerly known as SQL Server Data Services, provides a database in the cloud. As the figure suggests, this technology lets on-premises and cloud applications store and access data on Microsoft servers in Microsoft data centers. As with other cloud technologies, an organization pays only for what it uses, increasing and decreasing usage (and cost) as the organization’s needs change. Using a cloud database also allows converting what would be capital expenses, such as investments in disks and database management systems (DBMSs), into operating expenses. A primary goal of SQL Data Services is to be broadly accessible. Toward this end, it exposes both SOAP and RESTful interfaces, allowing its data to be accessed in various ways. And because this data is exposed through standard protocols, SQL Data Services can be used by applications on any kind of system—it’s not a Windows-only technology. Unlike the Windows Azure storage service, SQL Data Services is built on Microsoft SQL Server. Nonetheless, the service does not expose a traditional relational interface. Instead, SQL Data Services provides a hierarchical data model that doesn’t require a pre-defined schema. Each data item stored in this service is kept as a property with its own name, type, and value. To query this data, applications can use direct RESTful access or a language based on the C# syntax defined by Microsoft’s Language Integrated Query (LINQ).

There’s an obvious question here: Why not just offer SQL Server in the cloud? Why instead provide a cloud database service that uses an approach different from what most of us already know? One answer is that providing this slightly different set of services offers some advantages. SQL Data Services can provide better scalability, availability, and reliability than is possible by just running a relational DBMS in the cloud.

The way it organizes and retrieves data makes replication and load balancing easier and faster than with a traditional relational approach. Another advantage is that SQL Data Services doesn’t require customers to manage their own DBMS. Rather than worry about the mechanics, such as monitoring disk usage, servicing log files, and determining how many instances are required, a SQL Data Services customer can focus on what’s important: the data. And finally, Microsoft has announced plans to add more relational features to SQL Data Services. Expect its functionality to grow.