High Connectivity
Sunday, April 30th 2006 | Ismael Ghalimi
High Availability is defined as 99.999% uptime, or a downtime of less than 5 minutes and 14 seconds a year. Fault Tolerant is 99.99%, or 52 minutes and 24 seconds of downtime per year. Well Managed is 99.9%, or 8 hours, 44 minutes, and 9 seconds of downtime per year. Managed is 99%, or 3 days, 15 hours, 21 minutes, and 36 seconds of downtime per year. If we were to use similar metrics for measuring the connectivity of a mobile worker, Connected (99%) is what we could reasonably shoot for today, and Well Connected (99.9%) might become feasible once all airlines provide Internet access aboard aircraft.
In order to reach Connected level, one would need to be disconnected for no more than 87 hours a year, which is roughly the amount of time that a business user will spend in airplanes every year. If you find a way to get connected anytime you’re not in an airplane, you win. I am currently trying to each this level, and for me that means carrying a smartphone with EDGE Internet access (Cingular 8125) and a laptop with built-in WAN access (Sony VAIO VGN-TX790P).
Getting to the next level will require Internet connectivity onboard commercial aircraft. Some innovative airlines have started to provide Wi-Fi access to their passengers, but I do not expect it to become mainstream before 2010. In the meantime, Well Connected level will be the privilege of users who are fortunate enough to fly their own plane, or who do not have to fly at all. Make your pick!
Beyond this, Fault Tolerant or High Connectivity levels are past the point of diminishing returns. Unless you work as a bodyguard for the President of the United States, I do not know why you could not be disconnected for at least an hour a year, not to mention 5 minutes. Looking at the problem through that angle, it seems to me that EDGE, EVDO and 3G are getting really close to providing a fix for the most critical Office 2.0 bug: lack of connectivity.
Entry filed under: Office 2.0
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Ismael,
I need a clarification from you. If my recollection hasn’t failed me, Fault-tolerant is 99.999%, which is defined as no single point of failure. Every component, including the O/S, is at least duplexed, like a telecommunication network. In the 70’s, there were only 3 fault-tolerant platforms on the market; namely, Tandem, Stratus, and IBM System 88 (Stratus with IBM label). They were called the Rolls Royce of the computing platform!
High availability was a term coined back in the mid-80’s when DEC rolled out its clustering hardware and IBM released its 1st generation of Sysplex. I’ve been dealing with national support system engineers from Tandem, DEC, and IBM throughout my consulting years (25). Fault-tolerant computing platforms are more expensive than high availability platforms. For instance, Fault-tolerant VAX was more expensive than clustered VAX, with the same computing power but not the same realiability level. Who on earth would buy an expensive fault-tolerant platform when high availability (clustered) platform could provide a better up-time?
I believe that credit card companies, telcommunication providers, and banks still use Tandem and Stratus for their network management. When the configuration is at least duplexed, you can call them as “non-stop”, as trademarked by Tandem! Non-stop is almost 100%, as one dares to claim to achieve that all the time. One explosion of an atomic bomb can paralyze all computers, unless they are shielded like those in the NORAD command center, inside and deep down under a mountain!
Francis,
You’re absolutely right, and you should not read my post as anything more than a simple analogy. I am not trying to get high-availability of my mobile Internet connection. Instead, I’m trying to find ways to get connected most of the time. And for me, fault-tolerance is nothing more than carrying two communication devices (laptop and smartphone) and a pair of power supplies for each of them.
[…] What’s not working? Having 100 people and 45 demo pods connected to the Internet through one hotel connection while someone is trying to give a live demo is not a good idea. As was said before, Office 2.0 needs broadband connections to the Internet. If you’re offline, Office 2.0 does not work, period. This is one of the most popular objections to Office 2.0, but it’s also the easier to fix. All we have to do is wait. Within five years, anyone who will need or want to be connected will, anywhere in the world, at anytime, and at virtually no cost. If you do not believe me, try developing synchronization software that works, and let’s see who gets there first. And if you find yourself in a plane without any Internet connection, maybe it’s time for you to read a book or a magazine, or take a nap. Let’s get real: the need for high connectivity is an illusion. […]
[…] Let’s double click now, and focus on the implications for the Office 2.0 revolution that is so dear to our hearts — or mine at least. What the iPhone brings to the table is synchronization that actually works, for everything from contacts, to events, bookmarks, and pictures. That’s nice. But because you also get a full fledged Web browser backed with EDGE (and soon 3G, as Steve indicated) and Wi-Fi, you also have the ability to access all your data, always, from anywhere. All of a sudden, the need for high connectivity has been addressed, much sooner than anyone might have expected. […]
[…] I am not a strong advocate of synchronization, for I believe that it is too complex to use, and that it is solving a problem that will essentially disappear when most of the population gets high connectivity access to the Internet. The development of broadband connections and mesh networks will take us there sooner or later, and the question becomes what to do until this happens. […]
[…] I am not a strong advocate of synchronization… […]
Trackback this post | Subscribe to the comments via RSS Feed
Leave a Comment