Cloud Computing: Microsoft Joins Emerging Database-as-a-Service Market

It's a proverbial warning shot across the bow of "cloud computing" database service providers — most notably Amazon but also others. It's also an attempt to steam ahead of the likes of Oracle and IBM. But most of all, Microsoft's recent beta release of SQL Server Data Services (SSDS) provides an intriguing and potentially exciting opportunity for small and large businesses that are making a foray into the rapidly evolving Cloud Computing and Web 2.0 environment.

SSDS falls into the relatively new category of Database-as-a-Service” (DBaaS). Here is how it works. Let's say you're building a social Web site where visitors can post their high school prom photographs and share them with friends and relatives. On a conventional site you would need to procure and deploy a database to hold all the related data — names, addresses, photographs, reader comments etc. That could be expensive even without the ongoing maintenance and support costs. With SSDS, the site's database needed could be taken care of by the online SSDS Web data store. You provide the Web application and the data definitions, and Microsoft provides the Web-based database and the maintenance.

Analysts expect DBaaS to grow rapidly on pace with the cloud computing trend. The approach is expected to appeal particularly to small and midsized companies looking to serve widely distributed communities, but it also faces the inevitable security, availability and performance questions faced by software-as-a-service providers.

What's In a Name?

Internet Service Providers (ISPs) have been providing databases for years, so what’s new in DBaaS? The difference is that with the ISP model, your application is collocated with the database: in essence, you're simply moving the application deployment platform — the full stack — from your premises to the ISP’s premises. With DBaaS, the stack is dismantled, and your application does not (and typically will not) be collocated with the database. The database is truly a service, with all the benefits and drawbacks of a widely distributed service oriented architecture.

There's already plenty of DBaaS competition, with stalwarts as well as startups jostling for spotlight. Force.com, the Salesforce.com platform-as-a-service, includes a database component, as does LongJump, a start-up competitor to Salesforce.com. EnterpriseDB, another start-up, has an offering based on the open source PostgreSQL database and Amazon S3, and offered through Amazon’s portfolio of cloud computing services. Trackvia and Intuit QuickBase offer their own variants of cloud databases.

It's important to dispel any confusion between DBaaS and the closely related concepts of Data as a Service (DaaS), Storage as a service (to avoid confusion, let's not call it "SaaS"), and Information as a Service (IaaS). DaaS is the broadest concept, ranging from simple storage providers (such as GoogleBase and Amazon S3) and low-complexity database services (such as Amazon SimpleDB and Microsoft SSDS) to Web services providers (like StrikeIron, an IBM partner in supporting mashups) and value-added data services (such as Kognitio, which offers on-demand data analytics).

DBaaS focuses on one part of data services: The ability to support potentially complex data models by offering, quite literally, (relational) databases on the Web. Storage-as-a-service operates at a lower level and is the equivalent self-storage on the Web: You rent a volume of space and not much else is included by way of value-added features. Information as a Service (IaaS), like DaaS, is a more diffused and general-purpose concept. According to Forrester Research analyst Noel Yuhanna IaaS can be defined as “a comprehensive strategy for the delivery of information obtained from information services, following a consistent approach using SOA infrastructure and/or Internet standards such as RSS,” and the latest Forrester Wave report on the topic identifies BEA, IBM and Oracle as offering the strongest IaaS platforms.

Think of IaaS as SOA infrastructure with built-in information/data management capabilities and geared towards offering information services to applications. It's tempting to imagine IaaS as something that happens within the confines of an organization, but this is inaccurate since there is nothing that inherently excludes the use of cloud computing.

'Bin' There, Done That

At first sight, Microsoft SSDS seems to be going up against Amazon's S3 (Simple Storage Service) and SimpleDB. The former is nothing more than a flat object store — an uncomplicated storage bin without structure or rules. SimpleDB adds structure and metadata to S3 so you can define items (organized into domains); each item has a number of properties, and each property has one or more values. SimpleDB works with S3, so this is incremental (or differential) value for application developers.

Microsoft SSDS is similar to SimpleDB. You can define containers (think database tables, loosely speaking) containing entities (the rows in the table). Entities possess properties that have sets of name-value pairs. Two similar entities need not have the same set of properties; some books, for example, may have an associated movie name while others (that have not inspired movies) don't have to have this property. Security and billing for the database-as-a-service are organized through concepts such as customers and accounts.

Despite this competition and Microsoft's late arrival to the party, SSDS has an edge in its tie to the SQL Server platform — an edge others may find it hard to compete against, at least until the Oracles and IBMs decide to show up, too. Whereas cloud database providers like Amazon are largely limited to relatively simple data models (basic bin-like storage, simple name-value pairs, spreadsheet-like tables and so on), Microsoft has the might of the SQL Server platform behind its offering.

Old Wine in an Exciting New Bottle

DBaaS is indeed a step forward for Web 2.0, and the possibilities are exciting. SSDS, for example, is only the tip of the iceberg of what Microsoft could offer. Given its development portfolio, Microsoft could go beyond a full-service database solution — in itself a significant offering — into areas such as dashboards in the cloud or, throwing BizTalk into the mix, Processes as a Service. The possibilities for mashups and cloud computing solutions are endless, but will necessarily be tempered by the realities around business risk and value.

Not surprisingly, DBaaS faces several challenges, beginning with the angst over data privacy and security. Database (and hence data) availability and performance considerations come a close second; in these days of fierce competition and flighty customers, any downtime or perceived slowdown can lead to a lasting or even fatal loss of customer confidence. Finally, there are various operational difficulties working with DBaaS: defining and maintaining the data model, technology and application upgrades, integrating with non-collocated data sources and so on.

More surprisingly, DBaaS is already approaching commoditization in terms of capabilities, choice and pricing. Even as vendors are cautiously surveying their competitors pricing schemes, they are finding it difficult to diverge widely in their own — an indication, perhaps, of their inability to distinguish themselves from the crowd. One presumes that technology-rich companies like Microsoft will have a clear advantage in their ability to provide value-added features and services, but the winners will be the vendors that can demonstrate the sheer solidity of their offering in terms of data privacy/security, availability and performance.

A Bright Future

Things are going great, and they're only getting better
I'm doing all right, getting good grades
The future's so bright, I gotta wear shades
— Timbuk 3

As cloud computing picks up speed, DBaaS is expected to follow along. Analysts see it as a viable platform for small businesses and midsized organizations that may find it too expensive and burdensome to manage database environments. Forrester estimates the current DBaaS market size at $20 million and forecasts a rapid rise to around $400 million by 2012.

DBaaS will be most valuable for businesses seeking to deploy what might be called distributed communities — systems addressing widely distributed groups of employees, partners or customers using the Internet to reach wider audiences while controlling costs. These businesses will have to be willing to risk locating their data outside of their own physical control, although SaaS providers such as SaleForce.com have to some extent already led business down that intrepid path.

Businesses considering DBaaS should keep in mind that the concept is evolutionary not revolutionary, and is here to stay (and grow). DBaaS in its simplest forms will be increasingly commoditized and hence will remain affordable, but unanticipated or hidden costs — for example, related to manageability and a need for higher service levels — could be another story; security, availability, performance and manageability will be critical success factors. Businesses will be best served by setting a long-term plan for cloud computing, ensuring that the DBaaS vendor has immediate or promised capabilities to support the long-term growth path.