The age of online services

I recently switched to Linux. Since I have two laptops, I needed a way to share my personal email and RSS reader between them. After all kinds of complications and experiments, I ended up with Google’s Gmail and Reader. Add the other Google services I use regularly, Notebook and Blogger, and the ones I use occasionally, Analytics, Picasa and Docs. Taken together, Google gives me email, an RSS reader, a place to save links and a way to make things like documents, articles, presentations and photo albums.

Most of those services used to run locally in desktop applications. Google realized, like many others, that all of them could be offered online. It took that idea further than anyone else. Its approach suggests a set of common features for online services.

• 1. Access from a web browser with minimal requirements
• 2. Data sharing
• 3. Some ability to work offline
• 4. Integration with other services

Access from a web browser without special requirements is a basic feature of an online service. It makes full use of the all-in-one idea. The browser gives you everything you need to use the service, with no installation, upgrades or configuration. That makes it widely available without a special client program.

Google’s online services show why it’s important to keep browser requirements low. Every extra component you have to install brings problems ranging from security to interoperability. If your service requires Flash or Java on the client, it’s practically certain to be out of luck on mobile devices.

A service lets us share the data we manage with it. Since the data is shared through the service, a browser is all we need to reach it.

Some ability to run offline clashes a little today with keeping browser requirements down, though that may be less of a problem in the future. The Reader I mentioned, for instance, can be switched to offline mode so you can read your feeds with the browser disconnected from the internet.

Integration lets us combine services into larger systems or use them ad hoc. The resulting applications go beyond the scope of any one service. A calendar can work with email, for example, or serve as the foundation of a scheduling service of your own.

From a user’s perspective, online services have plenty of advantages and a few features to weigh carefully. The big change is that most of the risks sit with a provider whose infrastructure the user knows nothing about. The relationship rests on trust, sometimes backed by law or contract.

Someone else has access to data I own. I know nothing about how it’s secured, who else can get to it, or what processes and procedures protect it. Knowing what the data is and what it contains lets a provider target me, with ads for example. Connecting services makes single sign-on possible. A user who has authenticated to one service is treated as authenticated to others without proving their identity again. But breaking into one service can partially open the door to the others. Backups and failover are also the provider’s responsibility.

For providers, online services offer a great deal more control. A provider can maintain a service without much difficulty. It can also turn a service on or off selectively, for example to make it available only to paying customers. Providers can better observe and evaluate how people use a service, then improve it accordingly.

I think online services will slowly but surely displace the local applications we use for ordinary work and entertainment. Users and providers will both benefit. Let’s hope we don’t end up with a single company that has a near monopoly on most of those services…