IT Is Power. Try Living Without A Single Breadboard For A Day.

Don MacVittie

Subscribe to Don MacVittie: eMailAlertsEmail Alerts
Get Don MacVittie: homepageHomepage mobileMobile rssRSS facebookFacebook twitterTwitter linkedinLinkedIn


Related Topics: Continuous Integration

Blog Feed Post

Continuous Provisioning | @DevOpsSummit #DevOps #Microservices

The technology is available, be it Stacki with Salt, Ansible with Puppet, StackIQ Boss

Now, with more hardware!

September 21, 2015, Don MacVittie, Sr. Solutions Architect. The “continuous” trend is continuing (get it..?), and we’ll soon reach the peek of the hype cycle, with continuous everything. At the pinnacle of the hype cycle, do not be surprised to see DDOS attacks re-branded as “continuous penetration testing!” and a fee associated with receiving them.

But as with all hype cycles, there is some bit in them – usually quite a few bits in them – that actually pass the smell test in a real-world environment. With virtualization, we didn’t see the end of server farms, or a reduction in operations staff, or any of the more wild-eyed claims, but we did see increased server density per administrator, a result that careful observers had predicted pretty clearly.

And you can see that in continuous provisioning also. While the introductory blog – here – was pretty short and to the point, it does point out something obvious. For items like continuous integration to work, they must have a stable and dynamic foundation to run on. That is a pretty tall order.

Unless you’re making infrastructure – from provisioning to OS deployment – as repeatable and adaptable as Puppet or one of its competitors make application deployment and configuration, the continuous integration and continuous delivery have weak points at the infrastructure juncture.

ContinuousProvisioning

The Simple/traditional Continuous Delivery image, modified to insert provisioning.

And so it makes sense that a system to deploy and configure – hardware, virtual, and cloud – systems is integral to the other continuous buzzwords. Sure, if you’re 100% cloud hosted, a cloud-only solution works for this, but all three are listed because that’s what the market will demand. Even if individual products focus on one or two of these target platforms, enterprises – which live in a very real world of hybrid everything – will seek out ways to implement all three.

So to put meat around continuous provisioning, it would entail…

  • Ability to deploy machines without human intervention.
  • Ability to deploy to hardware, virtual, and/or cloud environments.
  • Ability to customize deployment per-server, and/or per-server type.
  • Support for common hardware/virtualization technologies.

For those not reading this on StackIQ.com, I do work for a company that (among other things) fulfills most of these requirements with both an open source product and a commercial product. And I could make this list a lot more detailed and focused, but that would increasingly make it seem like my employers’ products were the only ones that fit the bill – when in reality, IT just wants the ability to automate the entire process from bare metal or virtual metal all the way through functioning application delivering value. I wrote a bit about that here.

The technology is available, be it Stacki with Salt, Ansible with Puppet, StackIQ Boss, or any of a slew of other combinations. They still take a bit of work, but the combinations are getting easier by the day – It is a couple hours to go from nothing to fully configured Hadoop using StackIQ Boss, to use an example I’m deeply familiar with.

So standardization and uptake are what’s left between here and the end of this particular hype cycle. The desire to streamline provisioning is real and has been around for a very long time. The technology has finally gotten to the point of useful, and that leaves adoption. Adoption is coming pretty readily for the solutions that are ready for prime time.

And full stack automation will allow operations staff to work on higher-value projects than “configure the RAID, install the OS…” type work generally is. Higher value, and hopefully more interesting also.

More Stories By Don MacVittie

Don MacVittie is founder of Ingrained Technology, A technical advocacy and software development consultancy. He has experience in application development, architecture, infrastructure, technical writing,DevOps, and IT management. MacVittie holds a B.S. in Computer Science from Northern Michigan University, and an M.S. in Computer Science from Nova Southeastern University.