State paid $22K each for Internet routers

Amortized over 10 years, I would have chosen the 3945 everywhere versus sticking 2921s in some places (the ISR that would have been an alternative) and 3945s in others. Single Security Policy. Single IOS update Policy. Zero doubt as to what features will run in a particular location.

I think what most people have difficulty with is that they are comparing this decision to roll out a state communications infrastructure with the fact that they can go connect a $60 linksys wrt54G in their house and serve a dozen people without breaking a sweat. And get wireless as well! The issues involved in scaling that across the state, while looking to the future, and managing all that gear is a different challenge though.

I don't see any huge scandal here.

"Our main concerns were to not have something that would become obsolete in a couple of years," Gianato said. "Looking at how technology evolves, we wanted something that was scalable, expandable and viable, five to 10 years out. We wanted to make sure every place had the same opportunity across the state."

Wow. I've spent some time in West Virginia libraries, and yes the internet connection was slow, but it's not due to the router. And in most towns, libraries and schools are not going to be serving anyone outside their walls with some kind of WAN. Best case will be public wireless near the building.

And how is it that the "West Virginia Homeland Security chief" is "leading the state broadband project"? Seems like bureaucratic overreach, technical ignorance, and budget authority all wrapped in.

Why not spend $50 for a router for libraries and something less than $22k per school?

And then quietly replace it (after dumping the config and replicating the configuration) with a $100 half-depth Supermicro server bought off Ebay, running OpenBSD?

And then walk out, like the Grinch after he artfully puts Cindy Lou Who to bed, with a nice new 3945 for my own use?

EDIT: forgot to mention, I need a CF card reader also to get the config off the Cisco flash.

Earlier: "Gianato acknowledged that he didn't heed Dunlap's advice or wait for an evaluation.

'The routers already had been bid out,' Gianato said. 'I think John was looking at our needs now, not looking at our needs into the future."

---

So not only does Gianato admit he doesn't know anything (which means he's not even suffering from the Dunning-Kruger effect), but he also admits that he's not sure what his IT advisor was thinking about, but, eh, it doesn't really matter because the people who are the experts obviously don't think about the long-term and I'm not going to bother asking them about it anyway, because as we all know, bids are binding contracts. Right.

Out government should be small. It should perform a few basic functions like making sure we are secure, our trading interests are promoted and our banking and financial markets are healthy. It should educate our children to the best standards in the world. Finally it should make sure we are energy independent and our food and water is uncontaminated. That's pretty much it.

A little known fact in the tech community is that a $7,800 router provides exactly 16.016 times more opportunity to a student than a $487 router.

I guess someone involved in the decision got some kind of kickback, gift, or will have a nice cushy job offer waiting for them at Cisco when they get fired as they should be.

Remember this article next time some liberal ask to raise taxes and justifies this by saying how stimulus money is important, and why everyone should have the same opportunities (quoted in the article - as if a 22k router gave any kind of opportunity to people browsing the net in a public library in WV).

And cry when they get trashed at the end of their lifecycle without having ever been connected to optic fiber.

i have been - in my time (2000? 2001? 2002? 2003? 2004? ...) - involved into selling the national austrian television their first video streaming solution (think: free (as in free beer) windows media streaming server) - it took man-years. man-years cost money.

I'm not a programmer or technical person, I'm an end user who uses R, LaTeX and processing now and again. I hack an awesome spreadsheet. I have a degree and I do a professional job.

I could be a manager in a public service or an elected representative. Perhaps in a small rural authority.

Given the lack of consensus in this discussion thread, and evidence in the article of a similar lack of consensus between the various offices involved, how am I supposed to reach a decision?

Is there a case for some kind of planning toolkit or requirements estimation software? Is there an opportunity here?

PS: I and other colleagues did once have to help a senior manager spend out a £250k capital grant in 10 days. The idiotic spending deadline was due to delay in award of funds in a competitive funding round. We did ok but could have done better with another month or two to think through detailed requirements.

No you fool! These are designed to handle large loads, not protection from obsolescence!

I would still like to see a breakdown of the costs and what services are provided with the routers before I call it a total waste though. If it was just the routers then it was a waste. If it includes other things like a long term warranty & onsite service, then maybe not so much.

I remember a few weeks ago PB made a comment that at least social networks and alike make it easier to spread the news, which eventually will lead to greater good (http://news.ycombinator.com/item?id=3890543). At least that's true, but doesn't feel like it's enough.

No. This is not okay. No twisted sense of reason that only you have makes it okay. A $60 consumer router that needs to be upgraded every year is still cheaper than this. Even with support costs. People have been rolling out tiered infrastructure for years.

I imagine the government worker that signed off on this purchase worrying a lot more about a dollar increase in the local lunch special than millions spent on equipment.

In this story, at least 20 million of the money seems to be completely wasted. (Higher capacity router does not mean better internet access for 4-computer lib.)

Wait, not finished. So our corrupt official has 2 million in his hands. He may buy a car or a house, but probably he would spend it for something useful. So much of the value would return to the market. This is also true for the vendor company. But in the US case, the value is gone. it is like you burned the cash.

Moral of the story: stupidity is more dangerous than corruption. You may want to have a corrupt offical than a stupid one.

Finally, that stupid would never have survived in the relatively corrupt but wild political system of my country.

* Yes, I realize this would bankrupt half the military contractors as well.

Assuming that the government has some rules that requires them not to deploy soho or open source products and they have to spend the money on some big brand company: how come Juniper was not taken into consideration (or other known network vendors)?

I would argue that Juniper could provide cheaper equipment of simmilar of higher spec (let's ignore for a moment if that is oversized or not) - and it becomes even cheaper if you consider the simpler licensing terms and upgrade support.

Isn't the government supposed to run open tenders? Is cisco mandatory in the US?

Anyone familiar with telecom and living in PA, WV, other states that have Verizon as the ILEC, knows this.

A librarian will notice a computer missing, but an expensive router replaced with a cheap one won't even be noticed for years.

somehow i doubt these sites have multiple uplinks, so they certainly don't need BGP routers. even if they're doing IPSec between some locations this can be handled by an old piece-of-junk $100 computer running BSD or nix, e.g. with isakmpd and pf.

stuff like this is why america's infrastructure is crumbling: dodgy contractors looking for an excuse to upcharge municipalities for irrelevant gear and services.

There's a lot to be said for uniformity in deployment, but that's why you would define, say, three tiers of supported hardware instead of kitting out everyone with top-of-the-line.