Making the Cloud Safe for IT
Author: Ashish C. Morzaria   Time: 5:18 am           In: Conferences| Network
This article is part one of two in the series.
What happens when Web 2.0 technologies are so pervasive that everybody is online all the time and everywhere? At any given conference, you have people catching up on work (email), killing time (aka surfing), or working on a document. The bane of a laptop user’s existence at a conference is a) lack of wireless connectivity and b) lack of electric outlets to keep working throughout all those great sessions. The behavior of people trying to get a connection or simply some electricity is comical - until you become one of them.
(Technorati code: t2qfiymcj6)

So what happens when you have a Web 2.0 conference where people have laptops, netbooks, and iPods/iPhones everywhere? I’m attending the Enterprise 2.0 Conference this week in Boston and the organizers appear to have put some thought into this - the conference rate at the hotel includes wireless access (which typically costs $12.95/day) and many of the conference rooms have power bars under the tables to alleviate the need for attendees to cluster against the walls of each session room.
The common wisdom is that in the future WiFi will be ubiquitous and free - putting cellular providers and WiMAX proponents out of business. We’ve seen municipal networks, free WiFi at bookstores, cafes, and even campgrounds in Alaska (more on this later). A very large company I am familiar with has found network access so pervasive (even for those who travel a lot), that if your corporate laptop “expires”, you are encouraged to use connect to Windows Terminal Services (WTS) behind the firewall with your personal laptop to keep working until the company machine can be resurrected.
The subnet mask at the conference shows there are theoretically 1022 possible IP addresses, although it is hard to get one - even if your association to the access point holds for more than a couple of minutes. The only reliable way to access the Internet at this time is to tether a cellular device (Blackberry, iPhone) or cellular AirCard. However, this is only because the number of people who are doing this is small enough that network performance is acceptable. I would expect the same problem to happen if you have 100 people with AT&T 3G/HSPA cards within the same room. Very little else can be as frustrating as not being able to access an important online application when the person next to you is slamming out Twitter posts every 20 seconds.
For a conference that is focused on technologies that absolutely require network access, it is interesting to see people pull out a text editor or email client to continue their work (as I am doing now). Even at the enterprise level, companies have had to carefully plan their 802.11 networks for the same reason. There are not a lot of solutions to this problem - partly because until these online technologies get critical mass, there isn’t a compelling business case to solve it.
The ironic part is that network access is the antithesis of the “Network Effect” (quite literally) that social networking and online collaboration tools are trying to capitalize on - my network connection gets better as there are FEWER people using it. This is the same regardless of topology - during times of crisis cell phone networks are overloaded, companies see dramatic peaks in network usage as employees arrive in the morning, and even a home user running BitTorrent can quickly bring network performance to its knees for anyone unfortunate to be on the same pipe.
We have seen this problem before in the Cloud world - the need for multi-tenancy in hosted environments. While this problem hasn’t been completely solved, it is on the minds of every vendor looking to provide a service in the Cloud.
Today, the challenge for people accessing Web based applications is the performance, reliability, and availability of the server systems doing the hosting. Tomorrow, the barrier to adoption may be that users cannot reliably connect even if the server has 99.999% uptime.
In the meantime, we’ve solved the “lack of power sockets at a conference problem” - but perhaps we should have had each seat equipped with a Cat-5 drop to a network switch for the room. Could it be this simple, or is there another solution?
1 Response to The “Reverse Network Effect” - Part One
Chicken Alaska and The “Reverse Network Effect” (2/2) | Quest for the Cloud
July 1st, 2009 at 10:38 AM
[...] This is the second of a two part series. You can read the first part HERE. [...]