Mattress Shopping

Posted on April 8, 2013 by jcoehoorn

Something non-computer-related today.

My wife and I recently decided to spend some windfall money replacing our 30-year-old top mattress. Yeah. I wanted to record some of the things I found in shopping for the replacement. We settled pretty early on a memory foam mattress. These have amazing satisfaction ratings relative to the price. However, these mattresses also have a few drawbacks.

First of all, I need to give props to www.sleeplikethedead.com. It had a lot of good information, and it made the process easier for us. However, there were some things I learned that were not available there.

One drawback of memory foam mattresses is packaging. These mattresses generally ship compressed. A small but significant percentage of people reported that their new mattress did not fully decompress on unpacking. Reading a lot of reviews, I was able to discover that you can greatly reduce the chance of this by unpacking it as soon as it arrives. The longer it stays compressed, the greater the likelihood of a problem. Another issue is that the mattress is very heavy. A lot of people strongly recommend unpacking the new mattress directly in place, as it will be awkward to move later.

Another drawback is odor. A small but significant percentage of people complained of something called off-gassing, which is an initial chemical odor. This fades over time, almost always within a day or two. A number people expressed good results by unpacking the mattress in a garage or sunroom and allowing it to air out. This advise is at odds with that in the previous paragraph. Weighing the two, I recommend unpacking directly on the bed, but have a good fan (for ventilation) and some Febreeze handy, and expect to spend the first night, maybe even two or three, on the couch.

A third issue is firmness. It seems that different people often report different firmness levels, even for the same model mattress from the same manufacturing lot. After reading a lot of reviews, I have a theory for what is going on here. You won’t see this anywhere else (yet), but I think that memory foam reacts to heat. Different people put off different amounts of body heat when sleeping. Put off too much heat, and the memory foam layer will get too soft, compress completely, and you’ll end up sleeping almost directly on the support foam layer. Depending on the support foam and your body weight, this could result in a too-soft or too-firm sleep. Put off too little heat the memory foam may be too firm all by itself.

It used to be that memory foam mattresses also resulted in complaints of sleeping hot. This can still happen, but it looks like most mattresses made in the last couple years are able to account for this. It’s still worth bringing up, though, especially in light of the prior issue. You might still need a second sheet to help dissipate heat.

I also need to bring up durability. My research indicated most adults will need a 9 or 10 inch memory foam mattress. More is wasteful (unless you’re seriously overweight to the tune of 300lbs plus), and less is likely to result in early sagging, greatly reducing the life of the mattress. You really want about 3 inches of memory foam and about 6 inches of support foam. Even with this level of support, don’t expect your mattress to last like a traditional inner spring. You just won’t see 30 year-old memory foam mattresses out there. We were able to get a queen 10 inch mattress with a 20 year warranty for about $350, but I’m skeptical how much good the warranty will do us. What I’ve read is that the actual useful life of a memory foam mattress is much closer to 7 years. At $350, that means it works out to $50/yr, or just less than $5/mo for this thing, before buying new sheets. Cheaper mattresses will significantly reduce durability. Do check for your manufacturer on www.sleeplikethedead.com. This site has good information on the actual reported longevity from different manufacturers.

One other (minor) concern is sheets. You can’t walk into any old store and find sheets for a 10 inch thick mattress. We had to replace all of ours. Amazon has them for around $30, and we had budgeted for this, but it’s still something to keep in mind.

I would advise anyone else looking to make a similar purchase to be careful reading reviews online. Look for reviews from people who have owned the mattress for at least six months. Discount low-score reviews complaining about odor or firmness entirely. These are people who were surprised because they did not do their homework, or in some cases obviously have their mattress upside down (it makes a huge difference).

In conclusion,  I am very satisfied with our new mattress. It’s a huge improvement over the old one. In spite of this, after reflection I already have buyer’s remorse. If I had it to do again I’d do it differently, and I advise others to go this route as well. What I would do is find the absolute cheapest inner spring mattress that I could. Firmness doesn’t matter, because you won’t be sleeping on it directly. The purpose of this mattress is to act as a more durable support-foam layer. Then buy a nice memory foam topper. You’ll spend a little more initially, but you’ll make it back after a few years when you only need to replace the topper. You’ll end up with something that lasts a lot longer, and that is much easier to find sheets for.

Posted in non-computer, Uncategorized | Leave a comment

Simulated Dice Rolls with C# and Linq

Posted on March 22, 2013 by jcoehoorn

A common feature of certain table top role playing games is to have players roll dice for their character’s starting ability scores. Typically a score is the result of rolling three six-sided dice and ranges from 3 to 18. Quick: what is the expected outcome of a such a role? The answer here is fairly simple; most people know, or can intuit, that a single roll has an average of 3.5, so three of them would add to 10.5. That is the average stat score.

But let’s make this more complex. It’s also common for players to be allowed to roll four dice, and then throw away the lowest score. What now is the average stat score? If you know your probability and statitics or are will to lookup the formula you can still work this out in your head, but it’s a lot trickier, and not at all intuitive. In my case, somewhere along the way I forgot the math involved.

C# to the rescue! I don’t need to know the probability rules to find this answer. I can write a program to simulate a few million dice rolls and tell me the result. Of course, any language can do that. What makes this fun is writing each simulation run as a one-liner using linq. I’m not a fan of the query comprehension syntax, but I will make generous use of the linq extension methods.

Here’s the code:

var r = new Random();
int iterations = 10000000;
double result;
 
//roll three dice
 result = Enumerable.Range(0, iterations).Select(i => Enumerable.Range(0, 3).Select(o => r.Next(1,7)).Sum()).Sum() / (iterations * 3.0);
Console.WriteLine(result);
 
//roll 4 dice, drop the lowest
result = Enumerable.Range(0, iterations).Select(i => Enumerable.Range(0, 4).Select(n => r.Next(1,7)).OrderBy(n => n).Skip(1).Sum()).Sum() / (iterations * 3.0);
Console.WriteLine(result);

 

The first round verifies the 3.5 average per dice roll, and the second round discovers the new number when you discard the lowest of four rolls: it looks to be about 4. Be careful, though. It’s not exactly 4. If I add more iterations, the 3.5 number gets accurate quickly to more and more decimal places. The other number, though, seems to settle in somewhere around 4.08. This matters, because if you change the code to the total of all three rolls, rather than dividing by 3, it means you end up with 12.24 instead of just 12. Applying this difference to the original problem of ability score rolls, and if you want to know if your character is above or below average for a typical game with six abilities, the number to beat is now 73.5 instead of just 72. Add up a character’s level 1 abilities, and 73 and below is sub-par, while 74 and above is good.

The code itself is not that interesting. It’s worth pointing out here that the upper bound of the Random.Next() function is exclusive, but the lower bound is inclusive… so a dice roll from 1 to 6 means calling the function r.Next(1,7). But the main trick is the use of a projection (.Select()) on the Enumerable.Range() method to repeat the dice rolls. Once you understand how that works, the rest of it should make a lot of sense… if a bit cryptic. Again, this was an excercise to do this as a one-liner. In real production code I would likely use a regular for loop for at least the outer enumeration simply to improve clarity and read-ability. But I think I’d keep the inner Enumerable.Range() projection.

Posted in c#, development | Leave a comment

What I did Last Night

Posted on January 26, 2013 by jcoehoorn

Last night (or, rather, much too early this morning) I had the privilege to change out the old core network switch (a stacked pair of 3Com 4900SX’s) at the college where I work for a shiny new HP 5406zl. I started at about 2am, and as I sit in Starbucks typing this it’s now past 10am and I haven’t slept yet. I’m very tired, and so I’m going to tell this mostly with pictures.

The shot below is shortly after I began. At 2am there are still a lot of students using the network, but I have a lot to do in terms of setting things out so that he actual down-time is as short as possible.

5043_10151283919513402_1569393134_n

At last everything is ready. It’s almost 3:30, and the system still shows activity indicating too many students are awake, but I can get started anyway moving over administrative buildings… as long as I save the residence halls and server switch for last, I’ll be okay.

552995_10151283920088402_1086556007_n

Here’s a “before” shot of the wiring board. This is an area that’s been in desperate need of improvement for some time, so it’s nice to have an excuse to re-run all this.

154102_10151283918528402_1422702699_n

And the after photo. I could get things neater, but I’m also hoping to get a new rack for the servers this summer, and so anything more would be overkill right now.

424515_10151284239498402_1802183412_n

This is what $10,000 looks like. Actually, more, because there are some consulting dollars invested in this as well. The new switch is all up and running… well, almost. One fiber link is being stubborn, but we’ll work it out. If necessary, I should be able to bridge the link in question from another switch in the same building, but I don’t think it will come to that.

71808_10151284290728402_1350170431_n

 

I think it was about 7:00 when I took that picture. I had everything put together by around 5:30, but there were two configuration issues with the switch: there was not default route(!) and the ip helper entries for our dhcp server were not all set right. They were both the kind of things should be have been obvious and quick fixes, but when it’s 5:30am of an all-nighter, sometimes it takes a while to find that kind of thing.

Last up is the old equipment, waiting to be boxed up and stored against the unlikely event of a disaster that would take out its replacement. I think if we ever have to actually put this back in service I’ll probably go cry.

398118_10151284295528402_1602336082_n

Posted in networking | Leave a comment

Can we please get annual update rollups from Microsoft?

Posted on November 26, 2012 by jcoehoorn

One of the more tedious things we have to do in IT from time to time is apply Windows Updates to a fresh machine. Depending on your process and connection, it can literally take days to get a blank machine up to speed. Most of us have things in place to reduce that time, and we don’t have to babysit a machine through the entire process, but it would be nice to be able to reduce the turn-around time and complexity of preparing the new machine.

I think that Microsoft could create a lot of good will in the IT community by offering annual update rollups. Each January, take all of the security or critical updates released since the last service pack and put them into a single downloadable package. The package should be smart enough such that the new process for setting up a fresh computer is to apply the latest service pack (likely already part of the OS install), apply the latest update roll-up, and apply only the new patches released this year. I would use such a package even with machines that are relatively current.

I think part of the reason this has not been done in the past is the complexity. Do you include Office updates? What version of Internet Explorer? Media Player? Either the number of different packages gets very large and complicated, or the size of the package itself becomes unwieldy. Also, traditionally Microsoft has tied support longevity of an operating system to this kind of packaged update, though this is easy to address by specifically excluding these annual updates in the documentation that talks about the product life-cycle.

On the other hand, the benefits to consumers and IT staff are obvious, and I think there are some additional benefits for Microsoft itself that may address some of the challenges. What if Microsoft only supported the latest Internet Explorer, Media Center, and Office in the update rollups? What if there were a simple switch that allowed the package to update to the latest Internet Explorer and Media Center, so that it could apply those updates? This would be an easy way for Microsoft to help consumers and IT departments stay up to date on these products. It would do a lot towards pushing users from IE8 to IE9 and 10, for example. It would help reduce the fragmentation of their products and encourage users to upgrade to the latest and greatest, especially IT departments which tend to otherwise lag behind. This is the point that I think would justify the engineering resources from Microsoft’s perspective.

Unfortunately, at the moment I’m still just dreaming.

 

Posted in Windows | Leave a comment

Explaining Microsoft Surface Pro to Mac People

Posted on November 14, 2012 by jcoehoorn

Image you have a Macbook Air. But this Air is special. It has a full touch screen. More than that, the keyboard section can be easily removed and replaced. There is also a hardware button built into the screen section to bring up finder, with your Apple Store iOS apps featured. This device would be running an updated version of OS X, with modest improvements to make the OS work better in the case where the touch screen is your only mouse, and features to support tablet behaviors such as rotating the screen.

You could take this device, sit on your couch, lay the keyboard section on the coffee table, and use the screen just like an iPad. In fact, because it has a full x86 processor and RAM, this would be an iPad that is awesomely fast. When you’re done, you attach the keyboard half, put it in your bag, and head to work. At work, you open it up like a laptop, and the same device is a fully functional Macbook.

This is what Microsoft aspires to with the Surface Pro. It is significant product, and deserves some attention, not because it merely copies the iPad, but because it succeeds in moving the tablet platform forward, at least to some extent. It’s an iPad. It’s a Macbook Air. It’s both. Microsoft Surface has the potential to be the only computing device (outside of a phone) that many users will want or need, and that feat will be a challenge for the iPad to duplicate.

If you owned the Macbook Air described here, would you still want to buy an iPad? If you could get a Macbook Air and an iPad for the price of just the Air, would that grab your attention? Does never having to choose between iPad and Macbook when you leave home appeal to you?

Admittedly, this “iPad” would be a little thicker and about 1/2 lb heavier, but it’s still in the realm of comfortable couch use. It wouldn’t have all-day battery life, but it does still far exceed what we’ve seen from past laptops. And you’ll have to run Windows… but this isn’t your parent’s version of Windows anymore, either.

Posted in IT News, Windows | 2 Comments

One Simple Thing You can Start Doing Now with Windows 7 to make the switch to Windows 8 Easy

Posted on November 5, 2012 by jcoehoorn

The Start Menu has a been a fixture of Windows computers since way back in 1995. Windows 8 changes that in a big way. The Start Menu is gone, and there is no way to bring it back without the use of third-party software that is often quirky and not as well thought through as it could be. The transition to the new operating system can be confusing, but there is one thing you can start doing right now in Windows Vista and Windows 7 to prepare yourself for the changes and help ensure they don’t slow you down.

Windows Vista introduced a new search feature to the Start Menu. Open the Start Menu, start typing the name of the program you want to run in the search box at the bottom of the menu, and that item will appear after typically no more than the first 3 or 4 keystrokes. Bonus points if you can learn to do this without ever touching mouse; just hit the Windows Key on your keyboard and start typing, then hit enter when the item you want is at the top of the list. Don’t even wait for the menu to show; just treat the Windows Key as another keystroke. In no time, you’ll find this workflow is faster and easier than the old browse-the-menu-with-the-mouse scheme you had been using. This feature is carried into Windows 7 as well.

Here’s the trick: all of that still works in Windows 8. In fact, it works even better because of the way Microsoft handles the new menu: it’s faster, and your search will usually have better results. If you don’t have a touch screen, this is now by far the easiest way to find what you’re looking for on the start menu. It’s just that most people never learned how to use that feature on the older operating systems, and the visual cues that help you discover it are gone on the new operating system. However, if you start practicing this new workflow in Windows 7 or Windows Vista today, you’ll find it comes naturally for Windows 8… and I think you’ll find the transition to Windows 8 much easier.

 

Posted in Windows | Leave a comment

AppleTV vs AirServer in the Classroom

Posted on September 24, 2012 by jcoehoorn

AirPlay is fast becoming the go-to protocol for wireless display purposes. There is software to support mirroring your display to a device such as an AppleTV from iOS, OS X, PC, and even Android. In addition to the AppleTV, there are a number of other options to host a mirrored screen. This includes any device capable of running the excellent XMBC software, so dedicated hardware is as cheap as a $35 Raspberri Pi.

Recently I began to support a project that put a new iPad in the hands of about 35 college professors: nearly the entire faculty of the small college where I work. For early trials of the program, we put out a dozen real AppleTV devices and connected them to our classroom projectors, to support using the iPads in class. We are now removing these AppleTVs entirely, in favor of a product called AirServer.

We had a number of reasons for making the switch:

  • I had occasional reports of an AppleTV locking up after about 20 to 30 minutes of continuous use. The PC hardware running AirServer is easily capable running that long.
  • No reset or update issues. In order to prevent theft, we had to hide away the AppleTV units up in the plenum space in the ceiling, making resets difficult on faculty and IT both. This was especially trying when Apple released three software updates in four months. Updates require a click-through with the remote before allowing you to mirror a display, which effectively locked faculty out of the device.
  • No input selection issues. It’s hard to over-state this. We had a number of support issues when one faculty would set a projector or audio input to work with their iPad, and then forget to set it back. The next instructor in the room was not always tech savvy enough even to set it back. By using PC-based software, the workflow is consistent for everyone: log into the PC no matter what, and all audio and video run through there. This is a huge win for those faculty not yet using iPads, and also provided some benefits even to those who were, in the form of having the PC ready and warmed up in case they want to do anything else.
  • Easier to select the right room. With AppleTVs, every classroom was always listed in iOS as available for mirroring. With AirServer, the classroom does not appear until you activate it, making it much less likely to pick the wrong room. Yes, this does happen.
  • Easier Audio setup and support. AppleTV only supports hdmi or fiber audio, and neither was a good fit for our AV setup. This was a weakness of AirServer early on, but now that it’s fully supported AirServer wins here easily.
  • Cost. AirServer worked out to about $3 per classroom for us in purchase costs. That’s not a typo. When you add in cabling and audio converter boxes, AppleTV units were well over $150 each. AppleTVs were more work to set up and maintain as well.
  • Video support. Early on in testing, we were unable to play a simple YouTube over AirServer. For comparison, we tried over an AppleTV as well. We found at the time that the experience on both devices was horrible, but that the AppleTV was still better. This is no longer true. The experience is still less than ideal, but AirServer is now more reliable for this test than a real AppleTV device. In the end, if all you’re doing is showing a YouTube video, why not just use the PC? It’s only when you’re doing this as part of larger demonstration or app that it will matter.
  • Network setup. Because of the old wiring in our old buildings, the AppleTV units were relegated to wifi-only. This limited display quality. AirServer is able to take advantage of the wiring already in use by our classroom PCs, resulting in much-improved quality.
  • Support for multiple simultaneous devices. I haven’t seen anyone actually use this, but on paper AirServer supports up to four devices mirrored to the same screen. AppleTV does not.

I do have a few suggestions for anyone else working with AirServer. The first is to install Bonjour Print Services for Windows before installing AirServer. This was a key step in getting audio and YouTube working for us. It’s possible this is no longer necessary, now that AirServer is no longer in beta, but it certainly doesn’t hurt.

We also had to edit the Windows registry to get the bonjour name and activation to apply to all user profiles on a PC, by loading the hive for the default user. Reportedly this is fixed in the latest update, but it’s no longer a problem for us so I haven’t needed to find out.

We had to re-think our classroom network setup so that classroom computers appeared on the same network segment as wireless devices in the area, and re-work our wireless network setup so that we know for sure which network segment that will be. This was a big project that’s still not entirely closed.

Finally, we did not elect to have AirServer start automatically. The option seemed to cause reliability issues and make error messages show up at log in.

Overall, AirServer was a huge improvement over an AppleTV for classroom use. When we began the switch-over to AirServer from AppleTV, AirServer for Windows was still in beta, but now that it’s released I can recommend it whole-heartedly.

Posted in networking | 4 Comments

Are Client Access Licenses Enforceable?

Posted on January 5, 2012 by jcoehoorn

One of the less-fun aspects of managing Microsoft Windows servers is dealing with the Client Access Licenses (CALs).  CALs aren’t only expensive, but with different licensing models (per seat, per user, per machine, etc) and different versions (CALs bought for Server 2003 are no good for your beefy new 2008 R2 box) they are also complicated and easy to get wrong.  But sometimes I wonder: are CALs really necessary? What if we could just ignore them entirely? This isn’t as far-fetched as it sounds. I think if it ever really comes down to it, CALs might ultimately turn out to be just so much bad data. The reasoning goes something like this:

Software licenses depend on Copyright Law, which (in the U.S.) grants copyright owners certain very specific rights relative to their protected works… namely the right to say who is allowed to make copies, and when. Legal precedent and the U.S. Copyright Office have established several specific points relative to copyright and software: computer source code is eligible for copyright protection, compiled source code is still protected, and installing software from the internet or physical media to your hard drive copies the software. Moreover, merely running software that you’ve already installed involves making a copy from your hard drive to RAM and CPU. This means that software makers can require you to license, rather than buy, your software.  I don’t agree with everything here, but I’m okay with it… and I better be, because my opinion doesn’t matter. This is legal fact in most of the free world whether I agree with it or not.

Now let’s make the next leap to CALs. Imagine you license server software from Microsoft and are legally running it. You’re also running a legally licensed copy of Microsoft Windows (client) on another machine. You now connect this client system to your server, and exceed the numbers of CALs for that server. This is supposed to suddenly be a copyright violation. The thing is, I’m having trouble seeing what part of the protected software was copied with this action. Sure, a few bytes passed over the network, but these bytes are not part of the protected work itself. They were generated by the protected work, but at the express direction of the user. Microsoft or any other software maker will have a very difficult time placing a copyright claim on that data, and you can’t copyright behavior.  Most importantly, there’s no legal precedent or copyright ruling here that I can find one way or the other.

The significance is that running without CALs is probably not a copyright violation. However, it might turn out to be a contract violation. This is interesting, because contract issues don’t have the weight of the federal government behind them in the same way that copyright does. Just one example is that the DMCA’s anti-circumvention penalties would likely not apply. Microsoft can choose enforce these contracts, but that is a much more difficult proposition, considering the sheer size of their deployed base and the fact that few of these so-called contracts will even have a signature.

I need to stop and this point and mention that I’m not advocating that anyone stop tracking their CALs. I am not a lawyer, and I’m probably way off on some key point in my analysis. This issue is far from settled, and even if my thoughts on CALs here turn out to be accurate, Microsoft or the BSA could easily make your life very difficult and your wallet significantly lighter in the attempt to defend yourself. As a sys admin, I try to protect the organization I serve from that kind of trouble, and so I purchase and track the required CALs like everyone else.

But I can at least dream of the day when this is no longer necessary.

 

Posted in servers | Leave a comment

The Immutable String

Posted on August 4, 2011 by jcoehoorn

One theme I see often on StackOverflow is that someone will post a question related to strings, and someone or several someones else will come along and say, “You can’t; strings are immutable.”

This is all well and good and accurate, but I have a problem with it.  You see, I think most of the people who ask these questions have never or only rarely encountered the word “immutable” in the first place.  They will generally have a weak or vague idea what it really means.  I blame Eric Lippert.

Let’s get this out of the way: the root word of “immutable” is “mutate”.  Something that is immutable does not have the ability to mutate. It never changes.

Applied to strings, this means that a string instance, once created, cannot be modified.  If you want to change a string variable, you have to create a brand new string instance and assign the new instance back to your variable.

It’s good there are more programmers out there that know these $5 words, but it’s even better if there are more programmer out there that can communicate the ideas behind the words to others who don’t yet know them.

Posted in Uncategorized | Leave a comment

A Different Way to Wifi Part 2: Basics and Stages

Posted on April 8, 2011 by jcoehoorn

In my last post, I talked about what you need to get started building out your wifi network.  The main points were adequate switching and some basic network services like dhcp.  But the most important item from the list is a dedicated network team.  If you meet the first part, you probably have the second part to thank for that.  Unfortunately, while you might have some great people working on your network, if they’ve been around a while this whole “wifi” thing might be a bit scary for them.  Today’s post covers two areas.  First, a few basic things that will help your network guy put wireless in perspective.  Second is the different stages of your wifi deployment.

Basics

When we talk about wifi speeds, we tend to talk about them in the same terms as we do for wired networks.  This is misleading.  When we talk about a 100mbit or 1000mbit (1gbit) wired network, this usually refers to a switched network.   Wireless access points are not switched.

On a switched network, each packet moves between a computer and a port on a switch, or perhaps between two switches.  It’s the job of the switch to transmit a packet only when the line is clear, and only to the necessary ports.  This way, it’s like each computer is on it’s own private network.  You never have packet collisions and you leave as much of your network clear as possible.  This means you get the most throughput possible out of your network.

Two Network Packets about to collide

On an unswitched network, all packets go to all ports and visible to all computers.  If two computers want to send a packet at the same time, you have a collision.  Both computers must resend, and that particular bit of bandwidth used for the broken transmission is completely wasted.  This means that actual network throughput is much lower than the theoretical potential.  However, networks speeds are still advertised in these terms because there’s no good way to know how much lower the real value will be.   In a simple home network with few machines and little interference, you might get something close to what’s advertised.  More machines makes things worse quickly, and too many machines can even push this number all the way to zero.

The good news is that your network is only unswitched within individual access points, and there is some provision in wifi to allow multiple access points in the same space.   There are 11 available channels in the b/g wireless range, and more in the 5ghz range used by a and n radios.  The bad news is that these channels tend bleed into each other and overlap.  For example, in the b/g range you should really only use channels 1,6, and 11 most of the time (there are exceptions where you can use more).  That’s only three channels.  Since a single access point can generally only handle about 20 simultaneous clients, if you have more than 60 users in an area you could be in trouble.

There is one more important consideration to look at as your network grows: broadcasts.  Broadcast traffic can be really disruptive to wireless networks, and so it’s important to keep your broadcast domains small — typically no more than around 500 devices, though the explosion of mobile devices (that are on for shorter periods at a time) has perhaps pushed this number higher.  This is why you need equipment with good vlan support.  You will need to create multiple vlans (segmented by area) as your network grows beyond this number.

At York College, we are right on the cusp of needing an additional vlan.  We have well more than 500 devices.  Currently we have avoided this route, because many are devices are mobile (we have an iPod program) and we have some other things in place as incentives for students to use the wired connections available in their rooms.

Stages

When I think about stages of wifi deployment, I like to think in terms of three “C”s: Coverage, Capacity, and Clarity.

For the first stage, you are working just get a basic level of coverage to an area.  Imagine you are the only person on the network.  If you go to a specific location, would you be able to use the network there?  While building out coverage, you may find yourself tempted to look for stronger access points that can cover larger areas.  I urge you to avoid this, for reasons that will soon be clear.

Once coverage is established, you begin to worry more about capicity.  Sure, the access point on the floor below may be able to send a signal up a level, but if it typically already has too many users you may find you can connect but not actually get any work done.  It’s time to start adding more access points beyond the basic coverage level, because you need more Capacity.

As you add access points to meet your Capacity needs, you may find that in some places you have a Clarity problems.  Users at the edge of the range of an access point on a particular channel will send transmissions that reach area covered by the next access point over on the same channel, causing packet collisions and hurting throughput.  Again, you may be able to connect to your network, but find that you are unable to get any work done.  This problem is also called “Density”, but I find the term “Clarity” makes it easier to understand where we are in our deployment.

At this point, people tend to start worrying about things like rogue access points or interference from neighbors, cordless phones, and microwave ovens.  While that’s noble, it’s mostly a mistake.  The largest source of interference on your network is usually your own access points and devices.  To counter this, you want to look for access points that have smaller coverage areas and lower (or adjustable) power antennas, not larger coverage areas or more powerful antennas.  It may seem counter-intuitive, but this way you can have smaller size cells and you can put more access points in the same size space- you can have a more dense deployment where the frequencies you need stay clearer.  The downside is that this adds expense and management issues, but it’s how they do it with the big enterprise deployments, too.

Next time, we’ll cover site surveys and access point placement.

Posted in Uncategorized | Leave a comment