The world is awash with software with more coming every day. From mobile apps to enterprise systems, software is changing people’s lives in significant ways while being deployed at an accelerating pace.
More software is now being developed in so-called containers. Containers package up the services comprising an application and make them portable across different compute environments, for both dev/test and production use. This packaging allows containers to make efficient use of the underlying server infrastructure. With the proliferation of containers comes a container orchestration tool such as Kubernetes which helps manage tasks such as scheduling, load balancing, and distribution.
For many I&O teams, containers and Kubernetes is a new and different approach to deploying and managing infrastructure. Historically, infrastructure was deployed based on considerations such as performance, cost, reliability, scalability, etc. particularly for data centers. As the world migrates to public and private clouds, those considerations are shifting. For some I&O teams, deploying containers changes the game significantly.
I&O teams will need to work with development and dev-ops team much more frequently to get ahead of what’s coming, plan infrastructure driven by a new set of considerations and revise or replace I&O acquisition, deployment, and management processes.
Infrastructure and I&O Implications
There are several important implications that will impact I&O teams and an organization’s infrastructure.
- What are the security concerns/vulnerabilities with deploying containers and Kubernetes? The CISO, cyber, and I&O teams must be involved before deployment of externally facing servers/edge equipment. Without structured policies and testing to confirm that these (“servers” or Kubernetes and containers) are locked down, they will potentially create unnecessary security vulnerabilities that negate any increase in productivity that they have produced.
- Additionally, Kubernetes provide a means of creating software “on the fly,” therefore proper QA testing must occur. Proper QA testing will reduce cyber risk as well as minimize software quality issues or brand reputation problems.
- The majority of organizations, particularly larger ones that utilize containers and Kubernetes, will likely also have on-prem or cloud-based servers and databases. Of particular concern are the functionality, speed, and connection of databases to containers, especially mission-critical databases. These connection speeds must be evaluated and thoroughly tested before deployment.
MANAGEMENT AND OPERATIONS:
- I/O teams traditionally deploy software and hardware infrastructure independent of development teams. Deployment of containers and Kubernetes changes this since these are independent of hardware or operating system limitations. Physical, Cloud, and PaaS infrastructure all require IT staff experienced and specialized in the configuration of each platform. Kubernetes and container Software are micro-operating systems (think mini VM’s) that are prepackaged and pre-configured. They don’t require drivers and a new OS can be set up with the click of a mouse. These benefits may signal the end of proprietary hardware and operating systems, opening new opportunities for reduction in the IT investment and a drastic increase in agility, scalability, and time to market/deployment for application and software development teams.
- New processes for maintaining, monitoring, patching, updating, backing up infrastructure may be necessary, particularly for those with bi-modal IT.
- Since containers can only be deployed in the cloud – how will they mix with the on-prem and other existing infrastructure? The short answer is seamless performance may not be possible, at least for the near term. Currently, every major (and several minor) software companies (i.e., Google, Oracle, Microsoft, IBM, etc.) have developed platforms to manage and link Kubernetes with prem/cloud/ hybrid infrastructures. As more organizations explore and integrate containers and Kubernetes into their production environments, the management offerings will only get better, faster, and hopefully, less expensive.
- The I&O team will need someone with expertise with cloud, on-premise, containers, and Kubernetes. In the near term, organizations may find it difficult to locate talent quickly with those skills. Leaders will need to estimate their short, mid- and longer-term needs. If leaders prefer those skills be on staff, they must determine how to do so and in the expected timeframe (hire, retrain, etc.).
- Broad adoption of containers will impact the I&O team since traditional infrastructure specialists won’t be needed. Leaders must plan how to address their future talent needs with minimal disruption to their team and existing operations.
Considerations and Questions to Answer:
As companies deploy containers in private and public clouds, many will be operating bi-modal IT since some applications will continue to be housed in on-premise data centers. I&O teams will need to plan accordingly, even as they deploy infrastructure for containers.
Here are some questions for I&O teams to consider:
- How will containers connect to the Domain Controller / User Directory (i.e., Active Directory)?
- What must be done to plan for disaster recovery (e.g., back-ups), redundancy, failover, and overall image management?
- How will their mix of infrastructure vendors evolve in the short and longer-term? How will this evolution affect SLA’s and other performance considerations?
- How should infrastructure governance evolve to accommodate container and Kubernetes usage?
- As cloud usage flexes up and down, how will I&O teams respond to ensure high availability, performance, cybersecurity all while minimizing cloud operational costs?
- How will I&O leaders ensure they have the right mix of in-house and vendor expertise to respond to container and Kubernetes growth?