WAN services: In theory and in practice

Our recent discussion of frame relay over DSL gave a relatively strong endorsement of the technology.   In fact, even our discussion of mean-time-to-repair issues indicated that this shouldn’t be a major problem in most cases.

Unfortunately, not all cases turn out exactly as they should, and this is due to the human factor.

For example, we heard from a user who had just had a particularly bad experience.  He wrote, “The sad truth is that DSL – even business-class DSL – has much poorer SLA levels than does dedicated data.  Even worse, those SLAs are toothless because they usually specify loss of one month of DSL revenue if the vendor fails to meet the SLA.”

Our reader pointed out that the loss of $80 – the approximate monthly price of DSL service – doesn’t seem to motivate most phone companies a whole heck of a lot. 

“All of this would be hypothetical to me except that we have just gone through DSL hell with two offices,” our reader writes. “They’re small offices, so DSL is the right technology, but I’m not ready to convert all of my WAN to DSL.”

It appears that our reader’s service provider didn’t follow a “prior-proper-planning” mantra, and that the loop that was sold for DSL service actually was of marginal quality.  Further, there are ongoing problems with both enforcement of the SLA and finger-pointing between the DSL provider and the service provider supplying the copper to the DSL provider.

It’s tough to do a good implementation of bad technology.  But it’s really easy to do a bad implementation of good technology.

It is certainly possible for FRoDSL to be a good service for many of you.  But we’ll also admit that without an SLA that’s appropriately honored and a service provider that takes responsibility for the entire process, FRoDSL will work no better than any other technology.