As 2018 rolls in, we're here to make predictions for cloud platforms. Unfortunately, I'm not here to come up with a present of a crystal ball that predicts all the exciting news to come, so I'm a bad Santa.
However, I do want to discuss some major trends that are underway, and how they may become important in 2018.
- Multi-Cloud. Do you want to use multiple clouds based on each one being a best of breed, or perhaps as a specific way for hedging your bets? What are the costs when you realize each has a different set of APIs? Do you redo your work N times over?
- Can a common layer be inserted into the cloud platforms that provides a common layer -- such as a container orchestration platform or some application support layer? (Think of software systems such Apcera, or Mesosphere; in the container orchestration world, we see adoption of Kubernetes all over or could you even use just plain OSs such as Linux as a common infrastructure layer?)
- What are the implications for operations and security as you use multiple platforms? Do you train people N times over, and create N different run books?
There are a few networking issues related to cloud adoption. The location of workloads makes a difference. You can run them in a public cloud, but those are far from where the access (or data sources for IoT) is initiated from. So maybe running those workloads, such as analytics, closer to the edge may make sense, but now you revive the very data centers you were trying to get rid of. Speed of light, which determines latency in networks, is a nagging issue we can't get ignore.
Those are some thoughts and we'll explore them further in 2018.