My RSS Feed reader turned up two different articles the other day I thought I’d share here since they’ve been a topic of discussion– sometimes heated discussion– over the last six months. “Text, trust, and third parties” comes from Inside Higher Ed, and “Hopes for wireless cities fade as internet providers pull out” from the New York Times (you’ll probably need an account to read that one, but it’s free).
The first one about campuses outsourcing their email is especially interesting here since (as I was tipped off/told the other day by a EMUTalk.org commentator who has yet to verify their email address and thus has not had their comment posted) EMU has announced (in a forum many people don’t read at all) that the two choices that mysterious committee are considering are gmail and MeritMail supported by something called Zimbra.
Here’s that announcement, cut and pasted from EMU Today:
CAMPUS E-MAIL SELECTION COMMITTEE IDENTIFIES TWO PRODUCTS:
The campus-wide committee charged with developing a recommendation for a new university e-mail system has narrowed its consideration list to two products: Google Apps for Education (Gmail) and MeritMail (Zimbra). After significant research and testing, the following products have been eliminated from consideration: Windows Live, Novell Groupwise, and Microsoft Exchange.The committee, which includes students as well as faculty and staff from every division throughout the University, expects to conduct further research on both Gmail and Zimbra, and make a single recommendation to the CIO and the SOC sometime in April. The committee’s final report will include details of the process; the selection criteria; research, testing and customer reviews; the rationale for the recommendation, including usability, appeal, integration parameters, expected costs and infrastructure/technology requirements; a migration plan and timeline; recommendations regarding campus communication and training. Though some new policies may need to be developed, the selection committee has not been charged with policy development.
The overall goal is to have the new system ready for university-wide rollout in Fall 2008. If you would like to set up a free test account on either of the two products currently under consideration, please visit mail.google.com or www.zimbra.com. Test accounts on these systems may not have the full feature set that will be available to EMU. Please also remember that whatever system we select, EMU users will retain their @emich.edu e-mail addresses.
If you have any questions about the process or the progress of the committee, please contact selection_committee@emich.edu.
I can see the advantages of going with Merit and I can see there are some other major players who have gone the zimbra route– The Ohio State University for one. I still personally prefer Google, though that is mainly because I already have a Google account and such. I might try to monkey around with the Zimbra thing and see if we can’t generate a little discussion about that here some time next week.
Here’s a quote from the Inside Higher Ed article that kind of sums up what it’s about:
The possibility of outsourcing a college’s e-mail system has its perks, of course (including, in many cases, no advertising links at all for current students), but it has also raised some of the same concerns that accompanied Gmail’s initial launch. The question is whether, like the last time around, those concerns will fade away.
If e-mail lists and other online chatter are any indication, many college officials are increasingly worried about the legal implications of keeping all of their institutional e-mail data offsite. What would Google or Microsoft do if a court issued a subpoena for private data? Would they disclose students’ data in violation of the Family Educational Rights and Privacy Act? How would they respond to Freedom of Information Act requests for data from public institutions?
Anyway, the article goes back and forth on this issue. Now, I think these are interesting and important questions, but I still think we ought to outsource our email. Two passages from this article resonated with me on this:
“We’re not sure why people perceive Google to be different than” other providers, including those on campus, she continued (“she” being Theresa Rowe, the chief information officer at Oakland University in Rochester, which recently went with Google Apps for this) — a “perception that somehow our campus e-mail is private and outsourced e-mail is not.”
“I think sometimes they fear losing control,” she said, referring to college technology leaders. “[W]e’re replacing the technological controls with contractual controls and service-level standards, and that’s challenging, you really have to think that through.”
And then there is this from “Chris,” someone commenting on the article:
What strikes me when I read this article is how irrelevant this whole discussion is. The first thing I did when I came to my current institution is set up my University e-mail address to forward everything to my gmail account. Most folks I know (at least those with enough technical savvy to set up e-mail forwarding) do the same.
Things aren’t like they were ten or fifteen years ago when students got their first e-mail address when they showed up at college, incoming students already have e-mail, IM, a blog, a twitter feed, and a Facebook account. Most of their online communication is going to take place through systems and sites that have no connection to the college whatsoever. Universities need to be thinking about how to integrate their communications with the systems students are using already.
I don’t know who Chris is, but I think he/she has it exactly right: if campus users are already using all of these for-profit internet enterprises for various services, it just isn’t that easy or possible to separate a campus operated system from the rest of the Internet.
The second article from the NYT is about how some of the grand plans for municipal wifi services have been running into some significant problems. Here’s a passage from that article that I think kind of sums things up:
In Philadelphia, the agreement was that the city would provide free access to city utility poles for the mounting of routers; in return the Internet service provider would agree to build the infrastructure for 23 free hotspots and to provide inexpensive citywide residential service, including 25,000 special accounts that were even cheaper for lower-income households.
But soon it became clear that dependable reception required more routers than initially predicted, which drastically raised the cost of building the networks. Marketing was also slow to begin, so paid subscribers did not sign up in the numbers that providers initially hoped, Mr. Phillis said.
Prices for Internet service on the broader market also began dropping to a level that, while above what many poor people could afford, was below what municipal Wi-Fi providers were offering, so the companies had to lower their rates even further, making investment in infrastructure even more risky, he said.
The article also talks briefly about the technology that has been driving the Wireless Ypsi project too:
Meraki, a wireless networking company based in Mountain View, Calif., has jumped into the void in San Francisco with a program it calls “Free the Net.” The company sells low-cost equipment that can be placed in a person’s home to broadcast a wireless signal. The company also sells inexpensive repeaters that can be placed on rooftops or outside walls to spread the original customer’s signal farther. The combination of the two types of equipment creates a mesh of free wireless in neighborhoods. The company says it has almost 70,000 users throughout San Francisco.
So I think the moral of the story is that muni wifi has been tricky at a lot of places and the model that Steve Pirece et al have might be the one that works. I still personally have some questions on some of this stuff as far as how it will work in Ypsilanti in the short/long term though.
The last couple of times I have been at my favorite downtown Ypsi destination, Bombadill’s, I’ve played around a bit with Wireless Ypsi vs. the access I get from their wifi. It works, but I have two minor problems. First, in my experiences (and maybe this isn’t typical), it was dirt slow. Second, there’s an ad banner that goes across the top of my browser and that is always there as long as I am on the service. I guess someone has to pay the bills, whatever those bills might be, but this ad and the fact that it won’t go away after I’ve logged in and started using the service is kind of annoying.
We gave the folks at Bombadil’s a radio 41 days ago. For some reason they haven’t quite been able to install it. We have spoken to them a number of times and they keep telling us, really, this week. Once they do, Wireless Ypsi will be better in that part of town.
Give WirelessYpsi a try at one of the other downtown locations like Pub13, Tap Room, Biggies, or TC’s. If you are ever curious about speed go to to http://speakeasy.net/speedtest and then click on Chicago and see what the speed reports back. Let us know the results.
When I tested at Bombadil’s I got 800Kb down and 500Kb up which I thought was pretty good considering most free services at national chain stores or the airport are significantly slower.
As far as the ad bar, it is a requirement of the system from Meraki. Meraki puts up one ad, it is the local weather. That doesn’t seem too bad, you get to see the temperature outside. We can put up 5 additional ads. We use it to advertise local Ypsi business, special events, and ask for donations to Wireless Ypsi.
We could get rid of the bar but it would cost $100 more per radio from Meraki. They call it the Pro system. We already have 60 radios deployed across the city. That would mean an additional $6,000. Ouch. We fully expect nearly 100 radio before Summer starts. So that means an additional $10,000.
However, we can upgrade to Pro at any time and not have to replace radios. So if someone wants to give us $10,000 so they don’t have to see the ad bar, just let me know and I can swing by and pick up the check.
We make NO money from the ad bar and the ad bar installs no software, no CDO, and no browser plug-in. If you switch to a different network, just close browser and re-open and the bar is poof, gone.
We have had over 1,500 users since Jan 1, average 150 to 200 users a day, over 250GB of data transferred.
If we can get some contributions from the DDA, we plan to get the rest of downtown covered, get RAC, Riverside Park and the border streets to downtown covered. Hopefully we will know in April if they can help.
In the mean time, we are volunteering to help four other downtown districts in Michigan to deploy wireless networks and will hopefully have them going before the end of May. It is really cool advertising for Ypsi. Several of these communities have already sent reps to Ypsi to check out the network. They even stopped in to eat lunch and have coffee.
It is cool.
Cheers!
– Steve
That answers some questions– thanks, Steve. One other question and one suggestion:
* This is kind of a naive, techno/geeky question more than anything else, but side-by-side, is the transfer/speed rate of these Meraki devices the same as I would get from (for example) the wifi a place like Bombadill’s is going to provide? Or is it one of those things where the transfer rate is “close enough” but not quite the same, kind of like cable modem vs. DSL?
* Might I suggest for some of the banner things local bits of news/info about upcoming Ypsi events? Or is changing them on a regular basis more trouble than its worth?
The answer to your first question is “sort of depends.”
Sorry, but there are too many factors for an easy yes or no.
In a side by side comparison, we see the same if not better performance over the standard d-link, linksys, netgear radio. Most of those radios aren’t very good and are typically not configured correctly
The Meraki is not quite good as high end Cisco radios. But the Cisco radios can run from $500 to $2,000 depending on config. The Meraki’s top out at $100. So for the price, they pretty mush smoke everything else out there.
Now will the performance be less or greater for you when connecting to Bombadil’s system versus Wireless Ypsi. Once Bombadil’s plugs the Wireless Ypsi radio in, (please please please plug it in, it will really help the network on that side of the street) both networks will be plugged into the same connection. However. Wireless Ypsi could still be slower. Here is why.
With Wireless Ypsi, each connection is throttled. So depending on the connection speed to the Internet, the max speed you will see with Wireless Ypsi is 1.5MB down and 500Kb up. One of our partners has a 12MB connection, but you still get limited. That is a good thing because that means more people can use the system and each be assured of similar bandwidth rather than one person being a hog and ruining it for others. Another partner has 384KB up and down. So that means the Wireless Ypsi radio is slower at that location.
So if Bombadills has a connection that is faster than 1.5 down and 500 up, Wireless Ypsi will be slower than Bombadil’s. But only by that differential. But if Bombadil’s connection is less than 1.5MB down and 500Kb up, then WirelessYpsi will be the same as the Bombadil’s newtork. If I remember correctly, Bombadil’s has a 3Mb DSl connection so they get 3Mb down and about 384 to 512Kb up.
However frequent Bombadil’s users know the network is always going down as the router they are using is having some significant problems. So it is often getting rebooted two or three times a day. In that case Wireless Ypsi will be more reliable. For one the radios are much more stable and because they run at lower power are less susceptible to heat and fatigue problems.
Secondly because of the self healing nature of the WIrelessYpsi network, if one connection goes down, the radio automatically starts searching out for other gateways to connect. We are now starting to get good enough coverage in downtown that we can lose one Internet connection and no one notices the hole in the mesh.
Do internet connections go down. Yes, all the time. Especially low cost and residential systems.
As we grow the system we are starting to build some interesting data on reliability from the different ISP’s that are used. All the radios constantly self monitor and report back radio repeater or gateway problems. So we quickly know if there is a network problem.
In the some 75 days of the network, we have seen not one failure of the mesh, but individual ISP’s go down, Many for as little as 20 to 75 minutes. Others will go offline for 24 hours or more.
So given the nature of the mesh and distributed nature of the system, reliability of the network is running in the double 9′s (meaning 99.99% uptime) which for a wireless system is really incredible. By comparison, most wireless ISP run in the 96% percentile on a year by year basis. We sill need a full year under our belt to see where we are going but at the 75 day marker, the trend is VERY encouraging.
Another factor are the number of hops between radios before reaching a gateway. Also another factor are other systems in the area, Microwave ovens, portable phones and other devices that all use the same frequency. A huge factor is the quality of the radio in the laptop. Even same brand laptops will have different chipset for the radios. Also drivers are a real factor. yesterday someone was having troubles with teh Wireless Ypsi connection. They kept saying it worked great at home and it must be our network.
Problem was he had a really old Linksys router and home and an old drivers dating back to 2001 on his laptop. It did work and has worked for 6 years. But when he went out into the community it failed. Why? Because most systems he was connecting are the newer B/G radios with far different chipsets and signalling and protocols. So we quickly updated his drivers to the latest version. A vintage 2004 was still good enough and all of a sudden his performance doubled and reliability was perfect. he didn’t lose the connection once.
So there can be a variety of factors that can affect performance. It is why we continually monitor and test the system to insure optimal coverage and bandwidth utilization.
The benefits of Wireless Ypsi over a plain access point in a business is not only bandwidth throttling, but the client computers are isolated from each other so it makes it much more difficult to packet sniff the network And there is better security for the business that is connected to WirelessYpsi because their internal LAN is isolated. At some local coffee shops, I can see their print jobs as they go by, see their desktop and laptop computers with Quickbooks installed, and much more. Most small businesses offering free WiFi don’t truly understand how much data is ‘leaking’ on their network. If they did they would turn off the free WiFI.
That is why WirelessYpsi is an intriguing option for businesses that want o offer free WiFI to their customers, it helps to insure that bandwidth hogs don’t steal all the service, it has better security for each free customer and it is better secured when plugged in to the LAN than WiFi systems currently installed with open/free access for customers.
Regarding banners, you bet we plan to change them all the time, and we do change them to announce new events and happenings in the community. Anyone that is a partner in the network can get a banner into rotation and we will change them as as often as we can to keep them fresh and informative. There is no charge to a partner to put an ad into rotation. So this is perfect for places to announce upcoming shows, art show openings, special events and more.
What does it take to be a partner? Just buy a radio and plug it into your Internet connection and you are a partner.
Hope this helps. My offer to answer any questions is still open, you hadn’t called after the last discussion. If you don’t want to chat on the phone, next time you are downtown, let me know and we can meet for a coke or coffee and I can tell you more about the network.
Cheers!
– Steve