The new academic year began last week and the College is full of activity. The Clapp Library is looking busy and alive. One knows how busy the Library is just as you enter – the exterior door has some issue (wear and tear) and the next set of doors are typically left open (patrons are too busy to close them; they really think that they are swing doors that should close automatically, but they don’t!). Sorry, the doors are my pet peeve and my attempts to get them fixed permanently have always resulted in the answer of how costly it would be! I even suggested an old style cord of sorts for the interior door that the student or the staff at the service desk can pull to shut them. But then we have insurance liability to worry about because some may trip over the cord. All kidding aside, we are on a mission to get these fixed soon (Hope is everything)!

There are some days that you never forget – the landing on the moon (I was here listening to tamil translation of a BBC broadcast on radio), the Challenger disaster (Watching TV here ) and of course 9/11 (I was at work at Wesleyan where we all gathered to watch the horrible scenes on TV). I found this op-ed in NY Times on 9/11/2012 to be interesting (though we have heard many of these before, when you look at it collectively, it sheds a different light on the problem). Of course, this is just one version of the events preceding the horrible tragedy. Hearing some of the survivors of the tragedy recount their loved ones even now makes one feel so sad. On to other things.

We are in the process of developing a short term strategic plan. One question we are interested in the question is the future of networking. It is the most fundamental technology we rely on 24/7/365 (or 366 every four years) and take for granted. Whether it is cloud computing or Mobile First, all of which we will incorporate, having a reliable and fast network infrastructure is critical. Note that I said reliable first!

First things first – we need wired network regardless. Even the wireless access points or the cell towers eventually need the bandwidth that is available through fiber optic networks. If you are really interested in all the fancy things that the physicists and engineers have been able to do with high speed transmission through fiber optic cables and light, you can review it here. One of the most fascinating this for me is the WDM (Wavelength Division Multiplexing) method of splitting the light itself by wavelength and sending billions of coded data in parallel through the same cable. How cool is that! Secondly, there are many different types of wireless networks – Wireless PAN to WiMax to Cellular networks such as 3G and 4G LTE. I want to restrict the discussion below the the Wireless LANs, which is what we use at home or work to hop on to the network. Generally these use the protocol 802.11.

The convenience factor of Wireless is undeniable, however the speed of transmission is a limiting factor. There are several other important factors to consider in comparing wired vs wireless. Wired infrastructures cost considerably more. Right from the get go, you need to first decide where to place the wired connection in a room, run the wire and allocate a switch in the port that is closest to the office. Then these switches needed to be tied together. It is fair to say that the cost per port is around $250 depending on how costly the initial wiring labor cost is. Ongoing port cost would be approximately $100-$150 based on today’s prices and port speeds when you need to replace the switches. This cost has stayed roughly the same for the past several years because the connectivity speeds have increased, so if you calculate the cost per Mb of peak transmission rates, over the years we are getting a good deal. Of course, wired connectivity has the problem that it is highly restrictive.

The other secret that no one really wants to talk about is that we have overbuilt the wired infrastructure. In other words, the peak use of the internal wired network is typically in 20-25% range. With the exception of a very few end users, predominant user population does not need the capacity that we have built. Even if they do, it is for very short periods of time. However, you will hear about network capacity problem all around. This has all to do with insufficient network capacity going to the internet! That is the rate limiting step and not your internal transmission. Besides, there are many other contributing factors, a slow computer where some program(s) are chewing up the CPU, or a highly underconfigured server unable to keep up with the demands, or sometimes the multi-tier  security driven network architecture where some appliance in the middle is choking are serious contributing factors.

Wireless is not cheap either. You need to purchase the access points and you still need the wires from the access points to the switches. Since a single access point serves a large number of devices, you need far fewer network devices in total. What you gain is clearly the convenience factor for the end users. In terms of issues, wireless devices use radio frequency spectrum for transmission and the earlier protocols such as 802.11b and 802.11g used 2.4GHz frequency. Unfortunately, there is a huge problem in the allocation of radio frequency spectrum, which, in the United States, is the responsibility of the Federal Communications Commission (FCC). It is pretty messy. Not to mention that it is not enough for FCC to sort this out just here in the US, but it also needs to coordinate standards  with other countries, so that the devices that communicate can do so regardless of geographical boundaries.

It also turns out that 2.4 GHz is also the frequency used by many household appliances which interfered with wireless communication. I remember getting the looks from my wife when her internet connection severed everytime she went to warm something up in the Microwave oven (as you know, all IT issues at home turns out to be my problem!). But, it was my older son who claimed to have lost some good money in online poker because of my uncontrollable desire to warm up that south Indian dessert one day that sticks in my mind.

One of my good friends at Wesleyan also wanted to have a microwave oven in his class, running the entire 50 minutes as a way to get the students to be attentive and not wander off into the internet on their cell phones and laptops. I had to remind him that it is against the law to jam signals. In addition, metals reflect the signals, so libraries with metal shelves are notorious for poor wireless reception. But a larger problem was the interference problem from neighboring wireless access points, which can be terribly painful in densely populated areas and apartment buildings. You can read more about it here. Like with all technologies, solutions have been emerging including smarter access points that detect nearby transmitters and self adjust by shifting to a different frequency band within the allowed spectrum, or lowering the power to avoid interference etc. And it will only get better.

Whereas some of these issues have been addressed, including the shift to 5 GHz frequency to avoid interference, it is just a matter of time before other devices will begin transmitting at 5 GHz and we will face the same problem. This is a very good source if you want to read about the most common myths about wireless interference issues. On our campus, we have installed 802.11N access points which are capable of transmitting at a peak of 300 Mb. Of course, the devil is in the details. The actual throughput is a function of the number of machines accessing a single access point at the same time, how persistent the connections are (streaming media is pretty close to persistent, web applications are not – they are intermittent), how good a signal you have etc. etc. But, I would like to argue that for many of us, the wireless connectivity works. Pretty much that is what all of us use at home and multiple machines connect at home and since we moved to the web browser for significant amount of work we do, this works.

For example, my Communications, Administration and Planning group which moved to new offices in Clapp library, has gone fully wireless. Of course, we are not working with heavy multimedia editing and constantly saving those edited files on the network, but we are constantly connected. Most of our work involves the web browser with some local applications such as the Microsoft suite. Some of us work with images and video very often. At home, I have basically stopped watching TV except for some live sports. We watch a lot of things on YouTube. I have DVD player which connects to my wireless and then to YouTube, so I watch the YouTube videos on TV transmitted through the DVD player. Works fine for me! Of course, some claim that the fact that it works for me doesn’t mean it works for everyone.

But I am willing to bet that we all consume a lot of  multimedia rich content on the network in our homes using wireless and if that works, I think it is fair to say that most of our users should be able to work using the wireless infrastructure. However, wireless may not work for everyone or everywhere. For example – classrooms, public computing facilities, faculty and staff who are working on large multimedia files that they need to save on network drives at a higher speed – are clear cases where we need wired network connections. So, we will continue to have the blended network but we need to strategize on the shift from wired to wireless. And this shift is going to be based on user experience and cost considerations.

All CIOs have to assess at every point in time the answers to tough questions – if you have a dollar to spend (that is a big IF these days) and you have several options in front of you, you need to make an informed decision (sometimes not popular) on where you will spend the dollar. To me, this question applied to the Wired vs Wireless balance, is a much easier one to answer. If you have not already guessed the answer – here is a clue:  At Wellesley, starting last year, student residence halls have no wired connections (except for a few for administrative purposes).

But, it is also the case that a CIO does not decide on these unilaterally. We have a governance structure for that precise reason. I have to present the case for whatever my recommendations are and they help us chart the future direction. One thing is certain – I will continue to use wireless 🙂