Most of the focus of the DevOps movement is on the application development to deploy cycle, and that’s appropriate. When folks talk about the IT Ops side of DevOps, you’ll hear talk about IT process management tools like Puppet, Chef, and Ansible that are used to automate the deployment, update and remediation of software and VMs. Yet, DevOps is widely acknowledged to be more than just one process—it is a culture of collaboration, enabled by automation that results in high quality outcomes. As such, more IT disciplines are looking at how to apply DevOps principles to their domain. These domains include those that can be put under the umbrella of IT operations management (ITOM): network engineering and operations, disaster recovery, and infrastructure operations management. Many of these disciplines and organizations are currently highly siloed in traditional IT organizations, and tool suites are from an era that precedes the current emphasis on API automation. Some may wonder whether these IT arenas can adapt and adopt DevOps. The more pertinent question may be “can IT organizations afford not to?”
ITOM is a huge topic, and there’s no way to do an exhaustive treatment of how DevOps is impacting it in one article. However, we can at least take a look from a couple of angles. The first angle is to consider DevOps impact in terms of functional areas of operational management. ITOM can be roughly broken down into managing provisioning, capacity, availability and performance. Of these, provisioning is probably the area where DevOps has had the most impact. This only makes sense, since without automated provisioning of infrastructure and application components, it would be very hard to achieve the continuous processes that are a hallmark practice of DevOps teams. The trend of all IT infrastructure to become “software-defined” aligns with this provisioning automation movement.
Once we stray outside of provisioning, the reach of DevOps becomes more sparse. In the win column for DevOps is Application Performance Management (APM). Five years ago, there was a tremendous dissatisfaction level felt by engineers trying to understand application performance. So much so that a #monitoringsucks meme developed. Since then, APM has significantly modernized and moved profoundly towards DevOps-friendly approaches. Open Source Software (OSS) options such as the ELK stack (ElasticSearch + Logstash + Kibana) are seeing increasing adoption. Meanwhile, commercial SaaS offerings from New Relic, AppDynamics, BMC and Riverbed are offering nextgen, highly API-enabled options. Note that even though APM is considered a category of ITOM, it’s certainly the most dev-centric.
Infrastructure monitoring has lagged by comparison. The typical operations center arranges itself in a multi-level skills hierarchy. Practices are shaped around monolithic, GUI-driven tools, and manual processes are the norm. In fact, it’s fair to say that much of the institutional knowledge in many ITOM practices is literally “muscle memory.” That is to say, knowing where to click on the screen. Stovepipes of these hierarchies are formed around different types of infrastructure–compute, storage and network. Collaboration between stovepipes is the exception, and ‘hot potato’ troubleshooting routines are the rule. Rather than emanating from organizational DNA, collaboration often requires top-down pressure, usually to respond to crises. On the tools side, monolithic, enterprise software deployments and appliance-based architectures are common. These products have been primarily designed to deliver a GUI experience. Their APIs are usually second or third class citizens from the vendor point of view, and aren’t meant for end-users. Rather, APIs primarily exist for vendor to vendor partnership integration and high cost custom integration projects. Clearly, this picture is not DevOps nirvana—so how is DevOps impacting it? The answer is, so far, incrementally.
One of the clear signs that an IT arena is getting primed for DevOps is cloud and open source disruption, and by disruption, I mean going beyond narrow, point tools to general purpose platforms. Those same open source tools used for APM are making their way into systems management, and to a lesser degree in network management. We’re also starting to see “green shoots” of change in systems and network management tools, with new SaaS offerings, though we’re far from a tipping point in these segments. However, an interesting challenge for true conversion to cloud-based DevOps approaches is the need to address a Big Data problem. The amount of data generated by IT infrastructure and networking equipment is now so voluminous that traditional systems and approaches can’t keep pace. Legacy systems that have their architectural provenance in the 1990’s are based on scale-up models, and were never built for today’s data volumes. Even cloud-based solutions may not necessarily be scoped to address IT Big Data. Until ITOM faces this technology challenge broadly, it will remain an impediment to transformational change.
About the Author/Alex Henthorn-Iwane
Alex Henthorn-Iwane is the Vice President of Marketing at Kentik . Henthorn-Iwane has more than 20 years of experience bringing new technologies in networking, security and software to global markets. Henthorn-Iwane leads the global marketing strategy for Kentik and help the company bring its innovative story and solutions to network-dependent organizations around the world.
Prior to joining Kentik, Henthorn-Iwane was VP of Marketing at QualiSystems, where he oversaw worldwide marketing at the industry’s leading supplier of DevOps cloud orchestration software for hybrid IT infrastructure, and Packet Design Inc., a provider of network management software founded by serial entrepreneur Judy Estrin. Prior to that, he held product management roles at CoSine Communications, Corona Networks, Lucent Technologies and Livingston Enterprises.
Through his roles at QualiSystems, Packet Design, CoSine Communications, Corona Networks and Lucent Technologies he has acquired expertise in cloud computing and the opportunities presented through virtualization. He has written for Embedded Computing, Virtual Strategy Magazine, Datamation, SDN Central, Datacenter Knowledge and InformationWeek.
Connect with him on LinkedIn, Twitter, or SlideShare.