I wrote the title of this post at a bar in Monterey tonight, literally on a napkin (see picture), before walking back to my room. Earlier in the day I had taught a 3-hour workshop on technology planning. In both preparing for, and giving, the workshop, it was beginning to dawn on me the full force of things like Bitnami.org and Amazon’s Web Services. I mean, srlsy.
When I, as just a moderately savvy librarian, can learn maybe five to ten very specific steps and be able to deploy any application I would likely want to deploy, why do I need to talk to my system administrator ever again? Let alone bring this person pizza or cookies to keep them happy? Just asking.
So here’s the thing — here’s why I have the title for this post that I do. Let’s just say you need a full-featured web site. If it’s possible that some existing piece of software, such as WordPress, or Drupal, or a variety of other applications are what you need, then you are only a few clicks away from being up and running in the cloud. You need to understand this, and this is why the job of system administrator is in jeopardy — at least in libraries.
If you join that up with cloud-based offerings such as from my employer, or our competitors like Ex Libris or Innovative Interfaces, then you realize that the future is not in hosting your own systems, and employing your own system administrator — it’s in managing your own cloud-based solution. So that is why I am so down on the future of a library-based systems administrator. If you’re not at a large academic library, which will likely continue to need your skills, or some other large organization that has particular requirements not well served by cloud solutions, then yes, start job hunting. I mean, totally. You can thank me later.
About Roy Tennant
I agree that “the future is not in hosting your own systems.” Still, someone has to tend the library’s cloud garden. The Sysop title may fade away, but whoever it is will still have to do whatever it is. You may think setting up a site with Drupal is a snap. But if I say “Drupal” to anybody in my library, blank stares is what I get.
If you are a library administrator who thinks that cloud-based applications can be deployed “with a few clicks”, you are TOAST.
If you are a library administrator who thinks that cloud-based applications will run themselves once they have been deployed, you are TOAST.
Two words:
security
(You need to read it twice.)
Just because it’s in The Cloud doesn’t mean your data are safe, and that means someone needs to be on top of security updates. Either you’re going to become a 50% sysadmin babysitting your AWS machines, or you’re going to have to pay someone else to do so.
This isn’t to say I don’t think things like AWS and Bitnami are gamechangers. They are. But they only take care of hardware provisioning and basic installation, not hot-fixing yet another PHP bug at 3am. As the backend stuff gets commoditized, sysadmin functionality in relatively free-standing institutions like academic libraries will likely be pushed “up the stack” — focusing on cost models, maintenance, niggly interations between applications, privacy and security, etc. — and leave hard-drive crashes, backups, and network configuration to Someone Else.
So no innovation then either? Just take the meatball software on the cloud?
The definition of “sysadmin” may be a bit murky. So, some context…at my institution, sysadmins handle system provisioning, patching, security, networking — real “low level” stuff. They don’t like to touch “apps”. These “low level” jobs (minus the security of course!) are the kinds of jobs that are increasingly becoming commoditized by hosted services (or every locally with awesome systems like Puppet). And that’s great! It gives organizations a choice.
When you’re think about pushing systems outside of your own infrastructure you need to consider more than just the initial deployment. In fact, you mention that the future is about “managing your own cloud-based solution” and I couldn’t agree more. About the managing part, this is! Who will maintain the system? I’m not talking about keeping it running, that’s what the cloud service does. I’m talking about configuration.Who will integrate it with other existing systems? Who will even figure out if it can be integrated? Who will patch it, if you’re only using a hosted VPS? How will you test an upgrade on your VPS? How will you provision users since you probably can’t rely on an local authentication mechanisms (LDAP, Active Directory, etc). Who will be backing your data up? Do you really not care about a local backup for disaster recovery? For some of these questions there are particular cloud services that have answers. For others, there aren’t. For some questions there will be answers “coming soon”. But, if you’re not asking these kinds of questions you’re setting yourself up for some pain.
This is why, for me, the ease of deploying an application is only the tip of the iceberg. I’m not a sys admin, but I am an application administrator. I’m sure that at some institutions I would be considered a sysadmin. So, when you say that “I” (in these roles) am “toast”, I’m a bit skeptical. These are all serious questions and they will not likely be going away. Some might be commoditized, but others may remain unique to the institution/organization.
Is “the cloud” right for you or your system? It, as always, depends…
So, does this mean that OCLC, III, and Ex Libris are going on sysadmin hiring sprees? They’ll have to, in order to replace what libraries are getting locally. Furthermore they have to compete with Google, Amazon, Facebook, LinkedIn, etc. for talent since they have to hire in quantity.
Not so sure I would agree. I must admit that it is about 50/50 of institutions that consider cloud vs hosted internally.
Security, tight integration and control are some of the reasons they give.
One thing that the cloud offers is the opportunity to focus less on running network cable, figuring out power supplies, buying machines, plugging them in, setting up the host operating system … and more on configuring, customizing and creating applications that are of use in the library environment. I think we are seeing a transition of developer/sysadmin jobs into DevOps, where traditional roles of systems administrator and developer are blurred.
Good points, and I particularly like Ed’s formulation of “DevOps”, which seems to get at the particular set of duties a former SysOp might have when using cloud services.
Completely agree with Ed Summers here. I’m currently involved in implementing a ‘cloud’ based Ex Libris Primo installation, and even me and my team have trouble getting everything configured the way we want it, together with Ex Libris staff. Also, it’s quite difficult to find out what ‘we’ (a large complicated University Library) want.
But there is more. Not only developer/sysadmin jobs will merge, also information specialist/subject librarian tasks will be merged with developer/sysadmin tasks.
Great advice — there’s a dearth of technology implementation skills in libraries, so let the few remaining find new jobs!
The power and advantage of hosted computing environments and growth of ‘one-click-installs’ type solutions has not gone unnoticed and unexploited by library sysadmins. The reality is that SysAdmins have been actively rationalizing in-house infrastructure, helping move to SaaS and other shared computing environments etc. for a long time.
The future was never about “hosting your own systems, and employing your own system administrator” — that’s always been a means to an end. Further, we often work in larger institutional contexts that impact how, when and where we can engage with shared computing platforms.
But one difference is that when we make the same moves locally (shared services, shared computing, “internet-scale” etc.), that’s not “the cloud” — but when it’s a move to highly selected *marketed-as-cloud* services – then that’s “the cloud”.
There are many more significant vulnerabilities facing libraries that may result in toasting jobs, but the “cloud” is not one that sysadmins need to worry about.
Sorry, newbs– PCI compliance and the coming stringent data protection laws suggest just the opposite. Once we pass the same types of laws already in place in Europe (and already in committee in the Senate), sending off your patron data to a “trust me” vendor ain’t never gonna happen again. You can’t forget that there’s a whole world going on outside library-land, folks. Srsly.
Dont agree. We are facing a data deluge in our University. The system admin role will shift to new software and new system admin tasks: managing hunderds of Terabytes of data, grid/data computing, library domotica. Also inside the buildings system admins will get new tasks managing many electronic study rooms. Their role will be bigger not smaller.
I’m with Ed and Lukas – DevOps sounds like the perfect description for my current job. But can’t see the next step (merger with info specialist/subject librarian) happening without bigger cultural changes.
We’re already there — we have no system administrator, everything we do is hosted by third parties. This does not mean, however, that a librarian can just “set up Drupal” — that would unwise for a number of reasons (not least “What need are you trying to cover?” and “In what way does Drupal service the identified need?”).
In fact it means that I spend time (when not programming) in meetings and writing specifications about the “web appliances” that we rent.
An appliance in this sense is basically like a software fridge — you’re getting someone to provide you with a functioning CMS installation, a caching layer, or even an application server.
The point is, just moving your server to the cloud doesn’t remove the system administration work, it just removes the hardware aspect. A managed appliance, on the other hand does, leaving you with the responsibility of deploying content to these appliances.
As Lukas points out, this is more complex than it might seem because everything has to be specified. And you pay through the nose for it.
I would argue that a SysOp (Or SysAd) is an evolving job just like the role of a front-line librarian is an evolving job. The cloud and everything is great as long as everything is running smoothly and you are willing to settle for whatever comes out of the box (cloud). But when that box is full of …. or that cloud drops hail, then I’m glad there’s somebody that has a little more technical background to beat it into submission. Maybe too I’m used to our Automation Dept. doing more than just pushing buttons. They also take care of the hardware when it goes belly up, fill in in on our reference desk (both came out of our Adult Services Department), give Tech Time help programs and provide a whole range of services for us. I think the button pushers had better beware.
Hedging by noting that larger academic libraries still might need a sys admin was a good call, since from that vantage point “all cloud, all the time” sounds less than ideal. We already do many things using cloud or cloud-esque services, with mixed results. For one, there’s the inevitable network latency issues. There are times when that remote server feels very, very remote as the packets get slaughtered on their way home.
Beyond that, there’s the cost issue. I haven’t yet seen the evidence that outsourcing always makes sense. Sure, anything we can run on wordpress.com can be done cheaper with cloud solutions, but not so with repositories, the ILS, etc. The cost models there may look attractive on the surface, but the indirect costs make them just as expensive if not moreso than doing them in house. That’s not just a supposition; that’s my daily reality.
Well, it’s clear I struck a chord with this post and I appreciate the much more nuanced and thoughtful responses that my simplistic screed generated. I think I’m in agreement with the vast majority of you, in that although some functions of the SysAdmin at some organizations may be going away, the overall job simply changes to meet the different needs. Thank you for responding in such thoughtful and thought-provoking ways. There is much food for thought here.