Culture War: How Network vs. Cloud Impacts Tools
What Cloud Technology Means for Network Management
In my first post in this series I discussed the burgeoning cultural war between people and organizations that are cloud-centric and those that are network-centric. As I described, every so often a shift in technology is so fundamental that it both impacts and is impacted by the culture of the IT organization. I also pointed out that the organizations that are able to maximize the potential benefits of these shifts are the ones that come to grips with the cultural impact that these shifts create. In this post I'll discuss how the adoption of cloud computing impacts the way that network organizations should think about the management tools they use.
Any change as fundamental as the shift to cloud computing is accompanied by lots of hyperbole and statements that might come true in the long run but have virtually no impact in the short term. An example of that phenomena is the attention currently being paid to NoOps (no operations). A recent blog post pointed out that NoOps was being driven by the adoption of cloud computing and quoted Forrester's definition of NoOps as "the goal of completely automating the deployment, monitoring, and management of applications and the infrastructure on which they run." Looked at this way it is easy to see that NoOps is one possible end-state for the ongoing adoption of DevOps. It is also quite obvious that NoOps is an aspirational goal that will not have any significant impact on the choice of management tools for the foreseeable future.
To analyze how the adoption of cloud computing impacts management tools in the near term, we need to establish what that adoption does and does not mean. One thing it doesn't mean is a transition within a very brief period of time from the traditional model in which all IT resources are provided on site to a new model in which those resources are instead provided via the cloud. Over the last several years we've seen growing adoption of public cloud applications and services, but most large enterprises still provide the majority of their applications in house. In addition, whether driven by concern over security or performance, some applications will continue to be provided on site for the foreseeable future.
Another thing that the adoption of cloud computing doesn't mean is that cloud-based delivery will become the end-state for all applications. Choosing to run a given application over the public cloud doesn't lock you into always running that application in the cloud. Some companies find it advantageous to get started with running an application in the public cloud but to later bring that application on site due to economic, security, or performance considerations. The interest that companies have in being able to move applications back and forth was highlighted in the recent announcement made by VMware and Amazon.
The fact that application deployment will be diverse and dynamic — some will run on site, some will run in the cloud, and some will go back and forth — means that IT organizations must have management tools that are designed to perform well regardless of where the application is housed. Other key characteristics that IT organizations should look for in a management tool include:
- Easily deployed;
- Provides access to all available management data;
- Enables access by a wide range of users with differing access rights;
- Capable of sharing data programmatically;
- Economically feasible.
The above characteristics are very similar to the factors that are driving IT organizations to make increasing use of cloud services. It follows that IT organizations should give strong consideration to acquiring SaaS-based management tools. One of the many advantages of these tools is that they afford IT organizations the ability to try a tool with a minimum of overhead.
In my next post I'll explore how Kentik's approach to network management enables IT organizations to respond to the changes brought about by the adoption of cloud computing.