Information Technology Reference
In-Depth Information
ting up a new machine required the skill and training of a sysadmin. Now developers were
allocating virtual machines using API calls. Without the need for full-time sysadmins, de-
velopers were learning more and more system skills and bringing with them their penchant
to automate tasks. By making deployment and test coding part of the development cycle,
they created a new emphasis on repeatability that led to many of the techniques discussed
in this chapter.
Otherpractitionerscounterbysayingthatsomesysadminshavealwayshadanemphasis
onautomation, thoughoutside ofwebenvironments management didnotvalue suchskills.
Fromtheirperspective,DevOpswasdrivenbysysadminswhoconcentratedontheircoding
skills and began collaborating with the developers on deployment and code testing. Taking
these steps into the development cycle led to closer ties with developers and working as a
tight-knit group to accomplish the shared goal of increased uptime and bug-free deploys.
More cynically, DevOps can be viewed as system administration reinvented by sysadmins
who finally had management support to do it the right way.
8.3.2 Site Reliability Engineering
Around the same time, companies such as Google started being more open about their in-
ternet sysadmin practices. Google had evolved system administration into the concept of
a Site Reliability Engineer (SRE) by recognizing that all functions of system adminis-
tration, from capacity planning to security, were crucial to the reliability of a site. Since
2005,Google'sSREmodelhasorganizedthecompany'sdevelopersandoperationalengin-
eers to share responsibility for reliability and performance. The SRE model can be thought
of as DevOps at large scale: how do you empower 10,000 developers and 1000 SREs to
work together? First, each product or infrastructure component has a small team respons-
ible for it. For critical and highly visible systems, developers and SREs work together in
an arrangement that mirrors the DevOps model. Unfortunately, there are not enough SREs
to go around. Therefore the vast majority of these teams consist of developers using tools
developed by the SREs. The tools are engineered to make operations self-service for de-
velopers. This empowers developers to do their own operations. The SREs build tools that
arespecificallyengineeredtomakeiteasytoachievehigh-qualityresultswithoutadvanced
knowledge.
DevOpsisrapidlyexpandingfromanichetechniqueforrunningwebsitestosomething
thatcanalsobeappliedtoenterpriseandindustrialsystemadministration. Thereisnothing
uniquely web-centric about DevOps. Marc Andreessen famously said, “Software eats the
world” ( Anderson 2012 ) . As this happens, all facets of society will require well-run opera-
tions. Thus, DevOps will be applied to all facets of computing.
Search WWH ::




Custom Search