One of my recent posts was about a Poll I am doing on my website… Anyhow, the poll is done via an external provider – it is pretty neat, I like it very much, but when the service goes down, as it did today, it is a major issue:

It is an interesting dilemma. So as we enter the new era of consuming Web 2.0 — 3rd party services, we must also accept its limitations.

Typically this kind of issue, the minimum level of service to expect from a service provider, even for “free” services, is addressed via service level agreements (SLA). But, how to define SLAs for “free” services? What is the definition of an SLA for "free Web 2.0 services"? Can we expect a minimum SLA? It seems there should be one, since it is a symbiotic relationship – I separate important real state on the website or the mobile application for their product, and help create awareness about their product, and attract customers to their site, and to their ads, in exchange for their service, for their working service.

This issue is very applicable to mobility as well. As Mobility 2.0, or the convergence of mobility and services on the web continues to evolve, having services always available and functional is of great concern.

So yes, there should be a minimum SLA, and it is about minimizing hard failures as the one above —  when deploying network services, think failover, think high availability — make sure the failover support is there, and at least the fail-to server returns a friendly "Service Not Available" vs. a hard-error that breaks the customer, in my case, my page. Web 2.0 services is not only about the application or the service, but about ensuring the service is available when needed, which is all the time.