IBM hybrid cloud efforts revolve around a hefty dose of automation, an API ecosystem and using artificial intelligence to bridge clouds and applications.
Dinesh Nirmal, general manager of IBM Cloud Integration, outlined the strategy and multicloud landscape in a conversation with ZDNet.
Here are some of the highlights from our chat with Nirmal.
Automation in cloud and workflows. Nirmal said: "Every company is trying to automate their enterprise."
He added that AI will have a key part of the automation process, but there are other parts.
The biggest benefit that automation brings is it takes the complexity out and brings the simplicity in. I mean, you look at the API economy, right? That's what it brought forward is that now you can use APIs to code and make a simple code through API, and you can get your applications running. The third piece that we are really focused on, the whole hybrid cloud software strategy, is how do you unlock the value of data using AI, because that becomes really critical. How do you get insights into your data using AI?
Accuracy matters to a degree. Nirmal said every segment of an enterprise will have a different set of expectations for automation and modeling. He said:
When you do automation, there is automation within different segments. You can do automation within your whole line of business. The workflows are one, the business rules are one, or you can do automation within your data center, which is the IT operations. So where you have to make sure is that it's just like AI? There's prediction. There's different levels of prediction, right, that you can get a 70% accuracy. You can get 72% accuracy, or you can get very low accuracy, but how do you make sure the model is continuously learning to do AI? Same thing in an automation. How do we make sure we get to a level of accuracy around automation to make sure that the business process can get the work done that you want to get done?
And humans will have oversight of that automation. Nirmal said:
There is always going to be some level of human interaction, right? If I take a workflow, for example, it's not just if this process comes approved. These are multi joined, nested workflows. How do you make sure that that human interaction is still there, and we can get that work done? There is no elimination of human element, even in automation. It's more about augmentation of human element in there.
Hybrid cloud will be here for the foreseeable future. Nirmal said:
All customers or most enterprises want to go to a public cloud, but it's a journey, and they will be on prem or behind the private cloud for a very long time...
It's truly going to be a hybrid cloud for a very long time, and how does customer manage both? How do you develop once and deploy anywhere you want? Those are the kinds of challenges that enterprises will have to deal with. That's what we built cloud parts for on Red Hat, OpenShift to make it easier for them to manage their data, manage their applications, both behind the firewall and on public cloud, because it's truly containerized. It's truly microservice architecture. You develop behind your firewall. You can run it on your public cloud. You can use the same set of policies and rules in both places. So we really help our enterprises to move into that journey and manage their data and applications on both places, but it's a journey. It will take years for them to get to the desired state they want.
IBM vs. the public cloud giants. Nirmal said "we are in the second inning of the game."
If you look at our software strategy, we really want to be truly hybrid cloud. Why is Google and AWS coming to private cloud? Because most customers, most enterprises are there today. That's where their assets are. They have a journey to cloud, public cloud, but the reason AWS Outpost is happening is because you cannot win this game without a hybrid story. It's going to be a multicloud future.