As I write this, the major new organizations are reporting yet another BlackBerry service outage. The last one, about a year ago, should have served as a powerful warning that we can become too dependent upon systems that have inherent single points of failure. Such is not a good strategy, either for us users or our vendors. As wireless systems of all forms become ever more mission-critical, a fault-tolerant/redundant/etc. architecture and implementation become ever more important - again, both for us and our suppliers. You can see Network World's article on the current BlackBerry outage here.
One way to deal with this problem is to make sure we are not dependent upon any single element in the wireless value chain. Sure, we'd like to carry only one mobile subscriber unit, but why not make it two, even if the second is a bit crippled? That's what I do - a wireless PDA (a Motorola Q) and a simple flip phone with "Web" access. Both of my devices are on a single carrier's network, but that could be fixed. E-mail should be stored and managed on a separate e-mail service, and forwarded, if necessary, to the wireless e-mail service, like BlackBerry. I use 100% Web-based mail on all devices, which means that I can get to e-mail through any browser (OK, most browsers), even a stationary PC. The e-mail service itself is still a single point of failure, however; I'm working on that one, but a fundamental re-architecting of the whole concept of e-mail may be required to really solve this problem. But, in the interm, at least there are a few steps we can take to protect ourselves.
When I wrote about the last BlackBerry outage, I warned that suppliers simply must do better in the reliability department. We don't yet know the story as to what's the cause of the current outage, but we will shortly. And I'm willing to bet that RIM has still not learned its lessons here. As a consequence, we must.
Mathias is a principal at Farpoint Group, a wireless advisory firm in Ashland, Mass.
|
|
KISS Applies
For all our cries that e-mail is "mission critical" and our reliance on BlackBerries (and other hand held devices, including cell phones), there are some realities which argue for a different approach.
Reality 1: Internet mail is inherently unreliable, and becoming more so as more and more data moves across the network, and as more and more spam filters go into place. No one should be surprised when a mail message doesn't arrive - we should be grateful when it does.
Reality 2: Hand helds and cellular devices are garage door openers with screens and keyboards. They are susceptible to all sorts of problems, and the fancier they become (read that as the more features they support), the less likely they are to be able to work reliably.
Reality 3: Many people are lulled into a false sense of security, expecting systems to work at all times. For them, Plan B is to re-do Plan A.
The lesson is that what can fail will fail. It is the responsibility of the sender, not of the network, to insure that critical communications takes place. Certainly networks should be as reliable as we can make them for the price we are willing to pay, but expecting the Internet to be anything other than a wild west show or three-ring circus is to fail to understand the animal.
I Couldn't Agree More
Yes, we do take e-mail for granted. No, it doesn't always work. Yes, it is the sender's responsibility to make sure critical communications are completed. But, no, the industry has not done everything it can to make the Internet (and e-mail) reliable. My big issue is that one's e-mail service provider is a single point of failure; that's just the way e-mail is architected. But this could be fixed so that, for example, a redundant POP/IMAP/whatever site holds a message until it is certified as delivered to the end user. And e-mail service providers can do more to enhance the reliability, robustness, and fault-tolerance of their systems as well. Hey, this could even provide a leader here with some market advantage and justify a higher price.
Thx. Craig.
Not to distract from the
Not to distract from the otherwise sage advice regarding single points of failure... the poster references a CNN article as symptomatic of the Press remaining out of touch. The article {"BlackBerry outage widespread"} is dated Thursday, February 13, 2003. Hmmm...
Oops, My Fault...
That was an even earlier outage I'd forgotten about! I'll update the piece with a more recent reference. Thanks!
Craig.
Post new comment