ScienceLogic’s Erik Rudin on why businesses must take control of their infrastructure in a multi-cloud environment.
We’re in a period of huge disruption in technology. Just when businesses thought they had a handle on their IT estate, the cloud and shift to digital has come along and changed everything. Previously, businesses needed to invest huge sums in enterprise-ready infrastructure, but now it is available to everybody for pennies per hour. This has led to cloud-first companies coming from nowhere and disrupting long-established non-digital businesses.
Investment in the development and implementation of new infrastructure abounds. This means that the old ways of ensuring service assurance are no longer working, and manual processes have become unmanageable in a hybrid world where cloud instances are spun up and down at will – after all, there are no people in the cloud.
Within enterprise, the period of recent change has resulted in complex infrastructure environments that change almost daily. The latest research suggests that over 80 per cent of businesses now run a multi-cloud environment and the analyst firm 451 has released the results of its latest survey predicting that three in five enterprise workloads will run in the cloud by mid-2018.
Take control of now with a CMDB
A configuration management database (CMDB) is a huge asset, as it provides a complete view of all the IT assets in a business. It serves as the single, master list of the IT environment. It maps out and describes relationships between network infrastructure, systems, servers, applications and, now, virtualised hardware.
There are three core reasons for the CMDB:
- Change management: By documenting your environment, you understand the many interdependencies among the various components. This means that, when you change any component of that environment, you will have visibility on what other elements will be affected. This avoids a domino effect where a seemingly inconsequential update could catastrophically take the production environment offline.
- Incident management: When an unplanned interruption to a service occurs, the CMDB identifies other potential services that could be affected. For instance, if memory runs low on a server and causes a service to crash then other services running on that server can be checked. The CMDB will also help with root cause analysis, to understand the problem that caused the incident.
- Service Management – The CMDB creates a repository where individual assets (described as configuration items) can be grouped into a single service. Once a service has been created, all subsequent incidents and change requests can be tracked back to the parent service. This enables the business service owner to be notified of any interruptions or delays to service delivery.
However, the problem is that the CMDB hasn’t kept up in multi-cloud environments. Some analysts put the percentage of failed and out-of-date CMDB in organisations at 90 per cent. It’s no longer the golden source of data it once was.
Software Defined infrastructure and DevOps is breaking the CMDB
In a classical data centre scenario, the configuration items in the CMDB did not change very often. As new physical hardware was bought, it would be added to the database and as old hardware was retired, it would be marked as no longer in use. At a basic level, the CMDB was directly related to purchase orders made by the business.
The introduction of software defined infrastructure has changed all this. Compute, network, storage and other services on-demand mean that virtual hardware can easily be provisioned with very little effort. The ease of use of cloud computing has conversely made maintaining a CMDB very challenging.
Multi-cloud has compounded the issue. Best-of-breed approaches combined with decentralised IT decisions have resulted in many businesses now using multiple cloud providers on top of their classical data centre infrastructure. This means that keeping track of everything becomes even more important, but it’s also difficult. This is why so few companies have successfully compiled a complete, real-time view of their infrastructure.
DevOps and Ops need to work together to implement a successful CMDB
Alongside, and complementing, the increasing use of virtualisation, many businesses have moved to a new, agile way of working. By implementing DevOps methodologies, businesses are streamlining the deployment of applications and de-risking the process of releasing software by using continuous delivery.
However, classical operations teams still exist to support underlying infrastructure and these teams are increasingly becoming separated from the business functions they support. Businesses need to deconstruct these silos. If the DevOps team are spinning up production instances of virtualised hardware to support a new build of software, they need to ensure operations are aware or, preferably, involved.
Combined with automated discovery of infrastructure, the CMDB needs to be kept current and inconsistencies flagged and actioned. Once the configuration management database is up-to-date it will become useful to everyone. The more it is used, the better the information contained within.
With an accurate CMDB in place and a single service point for the estate, you can identify when you are over- and under-capacity, reduce the risk inherent in change management, and assist your efforts in automation. When things go wrong, you know who to call and have mapped the dependencies on that failure, which will reduce the time to resolution.
We are in a period of tremendous change and transformation in business, and to make the most of the benefits of a multi-cloud environment, businesses must stay in control of their infrastructure.