Hacker News Reader

Comment by i000 | Hacker News Reader

All University IT departments I have interacted with were terrible too. Universities simply cannot pay market-level salaries for IT staff because it would turn the whole pay-pyramid upside-down. The ones who stay, spend most of their time making sure their work is neither cheap nor good: You cannot use the cloud! but you can have this VMWare virtual from 2009. You need to store 500TB? we don't have it but don't you dare to use S3. To get stuff done you have to essentially run a shadow IT group. Let's see how the UCSF experiment turns out.

View on HN

They could pay what is justified if they wanted. University pay is usually topped by sports coaches (and not just head coaches, at many universities numerous assistant coaches make more than any executives at the school and all academic staff).

There is no reason IT staff couldn't be paid at market rates. Poor management of IT is due to poor management not because "Universities simply cannot pay market-level salaries for IT staff."

Right. There are nonsense "administrators" who make similar salaries to junior devs, but they can't afford market rates. I'm sure whoever is in charge of this bound-to-fail offshoring makes at least that much as well.

Paying market salaries for IT would mean that the people who generate the income for the university/medical system (football coaches through merchandise, doctors through billable services, faculty through grants and teaching) would often be paid less then the personnel they hire. No board of regents will approve this.

Which of course tells about the thinking of boards of regents.

Myself, I've often personally hired people who I have to pay more than I get paid.

"University pay is usually topped by sports coaches (and not just head coaches, at many universities numerous assistant coaches make more than any executives at the school and all academic staff)."

Coaches & sports staff at the high level you are referring to, are NOT funded by the university budget. They're funded by specific athletic funds that is from sports revenue & that boosters pay into.

This is not an equivalent scenario - you're basically saying that a University should go set up an IT Booster Fund and have interested parties pay into it in order to pay these market rates - which they wouldn't (would you? I sure wouldn't).

"The other problem with the salary comparison is that Alabama taxpayers aren't paying Mr. Saban, and so his salary doesn't take any money away from professors. One of the benefits to come out of the rampant commercialism of college athletics is that media conglomerates and sneaker companies are willing to pay huge sums for the broadcast and apparel rights. Thus, Mr. Saban will be paid out of Alabama's $70 million athletic budget, with little or no impact on academic departments."[0]

[0] http://web.archive.org/web/20160325043141/http://www.wsj.com...

But who pays for the infrastructure of the sports, from what I've seen of stadium and amenities investments it's almost always the public that pays for that, and in that case the state does pay a lot to the athletics funds. One have to be real careful when thinking that the abstract budgets are not related.

UCSF doesn't even have any sports teams...

Not all University IT is horrible, especially in areas with less tech overall. I've been at a few places on the West coast and in the Midwest. While it can be pretty hit or miss, overall, I'd say the Midwest schools have better IT. And it's pretty much why you'd expect - academia can't pay market level salaries in Silicon Valley. If you're good at running a data center, you're not going to do it for a university in California. If you're in the middle of Indiana though and want to play with high performance computing, there are only a few places to go. For example, Indiana University has always had top notch IT. They value that as a core strength of the University and it shows... their president is even an old head of IT.

A sincere question: Why have IT involved in those things? Just flood the place with wi-fi, put an RJ45 in every room, and good luck. That's how it is in my house, both personal and business use, and it works just fine.

Most of the staff are buying their own computers anyway, because academic employers don't pay for computers. My spouse won't let IT touch her computer because they screw it up. Her lab still uses thumb drives to move data between computers.

At the worst, IT creates hurdles like you describe. At best, a university does need some services, such as to let students sign up for courses, and maintain their records. But even those services could be outsourced. My daytime employer doesn't maintain those services at every work site, but either outsources them or manages them at a centralized location.

In a certain sense, a university in SF is like a person who wants to live in SF but isn't in a lucrative enough field to afford it. Such a person (e.g., a teacher or a baker) would simply leave town, but the university wants to maintain its presence.

Likely for security among other reasons. As soon as someone uses something like Dropbox with a weak password and it's hacked, then the university's data is out there or at a minimum it is perceived as the university leaking. Or a thumb drive with personal information of staff/students is lost and then leaked by the finder.

> Why have IT involved in those things?

It depends on the context of their use. For example, a neuro-imaging lab wants storage space and processing power. They need to store somewhere between 500GB and 50TB of data they'll gather over the course of a multi year experiment. They'll need a few powerful computers to run a data analysis pipeline on the data. None of the people involved are computer janitors, the best they can do is maybe set up their home wi-fi and write programs in Matlab and R.

As the lab PI, you have a few choices but also a few constraints. For one, the IRB told you that all of the data you generate has to stay "secured." You asked them what they meant by that and they told you that it had to be locked up in your lab. Does this mean the cloud is out? The IRB doesn't know what the cloud is, but "stored on some server in Ireland" sure doesn't sound like "locked up in your lab."

How do you solve the rest of your problems? Do you just go down to Best Buy and buy a bunch of hard drives? They don't sell a 50TB hard drive. How do I even make such a thing?

Absent an IT staff, you now have a PhD-MD and their grad students sitting around googling about RAID and Linux. They'll try to create some system on their own to hold scans of peoples brains and other kinds of PII and they'll try and make it all work in house because all IT does is flood the place with wi-fi and run the RJ45 in the room.

Or maybe the PI could hire his own IT company? Then they notice that the six other PIs in his building also hire their own IT companies and they start talking about pooling the costs...

... in many places you are not even allowed to setup your own wireless network and connect reauthorized equipment on the university one.

As well you shouldn't. Name any employer nowadays that would allow that.

That's the difference between "university IT" and "research computing". Those are typically two separate groups. Think of it as the difference between running the IT (helpdesk, support, backups) at a tech company vs. running a cloud datacenter. You need both, and they are both "technically" IT, but they require very different skill sets. What they are talking about here is pretty standard corporate IT. I'm not sure if the outsourcing contract involves the hospital's EMR or not, but for those numbers, I'd guess not.

In my experience, once you start using the terms "high performance" or "cluster" university IT starts to push the problem back on to the lab or over to research computing. If the university doesn't have a research computing group, then you have a whole lot of other issues.

There are still plenty of research computing staff at UCSF.

Edit: slight wording change about IT (see below)

"University IT" is more than a helpdesk, and also requires services like storage, backups, networking and monitoring.

It's very cost effective to host these in a university data centre. They probably get a good deal on power. The Internet connection costs a token amount at a few peering points. Petabytes of research data probably already needs a fancy backup system, adding some documents isn't going to increase the cost of that by much. There are already staff who deal with hardware.

I know this... I was just trying to make an analogy for others. (Poorly, it seems)

It's the same IT structure that you'd have at any major company of the same size. And even at that scale you have good IT departments and bad ones.

However, it can be rare for the University IT systems and research systems to share too much infrastructure, aside from maybe a building with common networking and power. For one thing, University IT typically supports Windows desktops and servers while research systems are by and large Linux/Unix based. They also have different security requirements. Research usually doesn't deal with PHI, and so doesn't have the same backup security concerns that IT does.

In the specific case of UCSF, they already made the decision to offsite all of their business systems (I think it was because of local power costs, seismic concerns, and lack of a modern datacenter). But they still maintain some research computing onsite, but it is not centralized.

What I described is taken directly from a scenario that happened to me personally while working in a neuro-imaging lab. The lab didn't have any research computing support, neither did their department, neither did their college. The university had an IT department.

The CS department had research computing support, but it was in a totally different college (and about a mile away on campus). In my story, university IT had solutions for the scenario I described.

And really, their computing requirements are far from a "research computing" group. You just need to store 50 TB of data ("I forgot how to count that low?!") and have a single 8 or 16 core computer to do some number crunching. The entire setup could (and did) fit in a half rack in the neuro-imaging lab (which is what they wound up doing).

But then the same limitations are passed on to the outsourced services, and it's even less flexible because it's not just a "because I say so", it's also a contract...

Reply on news.ycombinator.com