Design, deliver, and run enterprise blockchain workloads quickly and easily.
Predictive analytics in the multicloud
Cloud computing has plenty of complexities. And while many IT leaders would prefer a unified infrastructure, wherein the business standardizes on one or two cloud vendors, that is not going to happen in the real world.
The reason is simple: Applications the business depends on reside on a variety of clouds. Forcing users to stop using some applications and services in the interest of simplifying the company’s cloud mix is unreasonable. That means a multicloud strategy—managing multiple clouds simultaneously—is the only logical recourse.
Even so, managing the multicloud is a difficult task and fraught with often-unexpected obstacles. For example, abstracting the platform—simplifying the user interface by pushing complex details, such as computer code, to a lower level on the platform—is helpful for developers and users, but it can be more complicated for the IT operations staff. This sort of complexity increases management issues.
“When you move to the cloud, you add entire new layers of complexity to your management operations,” writes David Gewirtz on ZDNet. “Keep in mind the need to automate and manage across clouds, and plan and budget, not only for the resources provided by the various public cloud providers you choose, but also for all the software and licenses you'll need to manage it all.”
And that’s where predictive analytics comes in.
At the center of data gravity
Predictive analytics is a class of analytics that projects future events based on historical and real-time data patterns and trends. Predictive analytics are used today in myriad ways, including predicting and managing the cost of Amazon Web Services Spot instances, preventing server and network failures, and managing customer experiences.
The key to getting the most out of predictive analytics in a multicloud environment is in first understanding that data is at the center of everything. It is the gravity that holds all the business applications together. It is the driver of every business decision. It is the center spindle on which all analytics turn.
Use of these technologies can help IT predict, and thus manage, how each cloud service in the mix will—or won’t—be used over time; ensure licensing and service-level agreement requirements are or will be met; and predict bottlenecks and process entanglements, among other problem prediction and prevention functions.
For that reason, accessible and flexible data storage and a refined overall data strategy are essential to making the whole of business analytics work. Busting every data silo residing inside each cloud is certainly essential to making a collection of cloud services work holistically for the sake of the business. But that, in turn, requires a new analytics strategy.
“Now that multicloud consumption is the new IT standard, it’s more important than ever to have a strategy for analytics managing your expanding technical estate,” says Thomas Anderson, senior director of management software at Red Hat.
However, complexities in the multicloud may seriously strain many predictive analytics capabilities. Many new technologies are entering the cloud reality, such as artificial intelligence (AI) as a service, supercomputing as a service, data science as a service, data storage as a service, and the avalanche of Internet of Things (IoT) data. They all require more storage and cloud computing resources (even and especially in conjunction with edge computing).
Predictive analytics are part of the analytics mix for multicloud management. Among other things, it can predict resource consumption rates, costs, and availability; streamline internal processes; manage risks; power automation; share neural nets with AI; and even predict and direct the development of neural net algorithms for AI.
And that leads to predictive analytics used to manage infrastructure as well.
“As cloud and multicloud infrastructure becomes commoditized, making way for containers, functions as a service, and the new world of machine learning, it’s time for analytics and management to follow infrastructure,” says Anderson.
Predictive analytics’ roles in machine learning can enable companies to manage and optimize multiclouds, no matter how large and complex they grow.
Predictive analytics and machine learning share similar roots in neural nets. Deep learning is a massively complex function of extremely advanced algorithms. To make accurate predictions, the machine learning outputs are added to predictive analytics inputs.
As deep learning and AI progress, they will incorporate predictive analytics in their own algorithms—many of which will not be coded by humans. “In theory with AI, choosing the right mathematical algorithm becomes obsolete as the machine determines the right technique,” says Richard Boire in a post in Predictive Analytics Times. Expect more automation in cloud automation as AI learns to write its own code and select its own algorithms.
Developing a multicloud analytics strategy
Every company’s multicloud analytics strategy is unique. However, there are common things to consider in any plan to use analytics in the multicloud. Here are four of them:
Consider third-party multicloud monitoring platforms. “It’s no accident that Google acquired Stackdriver, which began life as a cross-platform monitoring platform. Google still pitches the value of multicloud monitoring as it seeks to compete more effectively with AWS Cloud and Microsoft Azure,” writes RedMonk founder James Governor.
If a third-party multicloud monitoring platform would not make things easier for your IT shop, your choices are to run analytics for each cloud service separately and aggregate the outputs in a report or dashboard. Or you could create your own analytics for cross-platform analysis, which—as with all custom applications—is rarely the ideal.
Adopt deep learning now. While still an immature technology, deep learning is already changing how business is done. However, it takes a ton of data and a lot of time to “train” a machine, so get started now. If you do, you’ll have additional insights to add to your predictive analytics inputs going forward, and you’ll be better positioned to focus on innovating, automating, and technology research as these technologies mature.
Use predictive analytics to contain data storage spread and costs. Storage costs are cheap, which prompts the use of the same datasets stored in each of multiple clouds. But that’s not really a cheap play; data management in terms of governance, compliance, data hygiene, and security are intensive and expensive. You don’t want to run analytics repeatedly on the same datasets in different locations either. Bust those silos, and make data accessible to and across clouds. Use predictive analytics to manage and optimize the processes and accessibility.
Plan for cloud and edge. Because IoT is making it more common to do analytics at the edge—close to the sensor, gateway, or user—more data and applications are also “out there.” Include edge management in your multicloud analytics strategy because at least some of that data, and nearly all the analytics outputs, are headed for the cloud at some point. Treat the edge as part of the multicloud—at least in terms of your overall analytics strategy.
Is this the end of the multicloud analytics story? Probably not. For one thing, AI and predictive analytics will likely disrupt the very notion of multicloud, as it is already doing in other domains.
However, the cloud is here to stay. Like anything else, though, it will morph over time. And as long as the cloud remains, so too will predictive analytics’ role in managing it.
You can expect AI to continue to gain strength in capabilities and popularity in analytics use. Theoretically, at least, AI should make managing a multicloud easier in the near term. The IoT and edge computing will also likely grow and tax cloud capabilities further.
Whatever the future holds, managing multicloud is an immediate business need. It’s best to put predictive analytics to work now to make sure your goals are met, processes are streamlined, and problems are prevented. Plus, predictive analytics can give you a much-needed heads-up on what’s coming next.
Predictive analytics in the multicloud: Lessons for leaders
- Multicloud refers to the mix of cloud services and providers your company is using. Complexity is high and management is a challenge.
- The multicloud is growing as more cloud services are added, such as AI as a service, supercomputing as a service, data science as a service, data storage as a service, and the avalanche of IoT data.
- Predictive analytics allow IT to accurately predict resource consumption rates, costs, and availability; streamline internal processes; manage risks; manage licensing; and power automation across cloud services and providers.
This article/content was written by the individual writer identified and does not necessarily reflect the view of Hewlett Packard Enterprise Company.