[Baylisa] The Future of Systems Administrators

Sat Jul 18 12:14:51 PDT 2015

My Devops friends tell me the process is devops, while the job description is SRE. When LISA first got started in the early 90s we worked hard to define various grades of sysadmins based on their skill level, experience etc. Would be nice to revisit that process if it hasn't been done already.
You're exactly right regarding the increasing concentration of physical hardware in the hands of fewer players. It isn't just sysadmins that are affected, its the various hardware vendors as well! Regards,Sandeep Cariapa
      From: Roy Rapoport <rsr at inorganic.org>
 To: Sean Hart <boardnutz at blacklight.net>; baylisa at baylisa.org 
 Sent: Saturday, January 17, 2015 11:59 PM
 Subject: Re: [Baylisa] The Future of Systems Administrators

On 1/16/15 6:14 PM, Sean Hart wrote:
> So, first off.  A lot of the low level sysadmins (kernel hackers,
> rack-and-stackers, killer network folks, etc.) will go to huge
> installations like Rackspace and Joyent and AWS.  They will do awesome
> things there.

Putting aside the description of rack-and-stackers as doing "awesome 
things," I don't think it's actually accurate to say "a lot of the low 
level" folks will end up going to Joyent, Rackspace, etc, because with 
the increasing concentration of physical hardware in the hands of 
relatively few players, the number of jobs dealing with hardware doesn't 
increase -- it decreases for exactly the same sizing efficiencies that 
we as customers then end up taking advantage of.

> The Systems folks who do software installation and configuration/tuning
> will pick up chef/puppet/ansible/$CONFIG_MGMT_OF_THE_WEEK, and
> everything will be represented as code changes.

Some, certainly.  I brought Puppet into Netflix (and some other people 
later brought Chef in); but it still wasn't enough to save IT from being 
considered obsolete to production management, partially because we had 
so many IT people mired in an old and systems-specific way of thinking 
about what we built, and what the definition of "success" was.  And that 
old way is basically bankrupt -- IT has traditionally done a terrible 
job serving its customers, I'd argue (though a fair case could be made 
that IT's typically put in an almost impossible position where it's very 
unlikely to do anything but a terrible job -- one of the reasons I'd 
likely retire from the workforce rather than ever go into IT again).

> The all around, jack of all trades folks...  There's a lot here, but
> will be mostly writing automation code of one form or another
> (Provisioning, install, config, deployment, continuous
> integration/deployment, etc.).  They will be working on the
> Infrastructure as code movement and moving further and further up the
> stack.

We've been hearing about "moving up the stack" in various contexts; 
remember when we heard about this in the context of the offshoring 
movement and what we as engineers needed to do to stay relevant and 
employable? We were all going to become architects, and product 
managers, and ... a whole bunch of other positions of which there 
weren't enough for everyone :)

-roy

_______________________________________________
Baylisa mailing list
Baylisa at baylisa.org
http://www.baylisa.org/mailman/listinfo/baylisa