The Myth Of Best Practice


It’s yet another marketing term invented by sales to differentiate one’s product or service in the marketplace. Everyone wants to say their way of doing things is better than everyone else’s. But with no objective standard (and no industry consensus), how can one make such a claim? Being an equipment manufacturer doesn’t make you an expert. You may know your firewall better than anyone else, but it doesn’t mean your ideas about the deployment and operation of firewalls are best. There are many different environments, and you are likely to not have experienced many of them.

Every design incorporates many requirements that often are not explicitly stated ,and every design involves a series of trade-offs. The relative importance of each of these requirements and trade-offs vary from design to design, and can even vary over the life of a single project.

Here are just a few examples:

  • equipment cost or budget
  • scheduling/time to deploy
  • space/power requirements
  • compatibility with existing equipment
  • equipment availability/delivery times
  • ease of use/end user training cost
  • administrative and maintenance costs
  • technical skills of administrative staff
  • relative rate of change
  • industry competition
  • hours of operation (9-to-5 vs 24×7)
  • security requirements (a church might have different security needs than a bank)
  • reliability/availability (the elusive five 9’s)
  • compatibility with existing operations and practices (end user and administrative)
  • availability/cost of telecom circuits or fiber optic cable
  • legal, contractual or regulatory requirements

Every organization ranks these items differently. Some might consider cost to be the most important. Others might consider user experience to be paramount. A client may think hardware redundancy is important, but not if it costs more than X. An MPLS-based network might make sense, but not if the staff isn’t trained to maintain it. Good design is often a matter of compromise. It’s like the old saw: “you can have it fast, cheap and good. Pick any two.”

In any design, there are many factors to be taken into consideration. A cookbook design (Cisco’s SAFE blueprint, for example) makes certain assumptions in an idealized world. Few organizations have the same set of assumptions.

So even though I’ve designed many Internet perimeters, each one is unique, because each customer has different requirements and makes different tradeoffs.

Chasing after best practice can immobilize clients. Until they agree on a best practice they are loathe to make any improvements and risk being something less than best. Often, they end up doing nothing.

So rather than promoting the idea of a single “best” practice, I encourage my clients to think of “good practices.” There are many good practices, and the more good practices they adopt, the better off they will be. Conversely, there certainly are bad practices to avoid, and it is better to enumerate some of those rather than concentrating on reaching some impossible ideal. Often, just advising a client to “stop doing Y” immediately results in tangible improvements.

One response to “The Myth Of Best Practice

  1. "Best Practice" is a marketing term. I’m glad to see someone finally stated the obvious. However, I opine that the only true "Best Practice" would be to use Ethernet instead of Token Ring or AppleTalk; better yet, Twisted-Pair instead of Co-axial.

Leave a Reply