jro-grafik - Fotolia
For reasons ranging from integration to management, enterprises and cloud services don't always mix well.
"There are ugly things that have to happen in order to make the enterprise work and play with the cloud," says David Linthicum, SVP of Cloud Technology Partners, a cloud consulting firm based in Boston, in a recent podcast.
To help enterprises address these challenges, public cloud vendors need to place more of an emphasis on hybrid cloud. In this podcast, Linthicum and Sandeepan Banerjee, SVP of engineering and operations for container management company ClusterHQ and former Data Head at Google, discuss this topic, as well as new cloud technologies, such as containers and machine learning.
Will public cloud vendors fill in container ecosystem gaps?
Containers are one of several new cloud technologies making their way into the enterprise.
"Google has famously said that everything inside Google runs on containers … in that sense, everybody who is running on Google's infrastructure today is involved in a container deployment," Banerjee says.
With the increasing popularity of containers, new tools have emerged, but there are still voids in the container technology ecosystem, specifically for database monitoring management, security and governance -- which are important capabilities for the enterprise, Linthicum says. It will take some time for public cloud vendors to fill in these gaps.
"I think everybody is looking for Docker, CoreOS, Google, Microsoft and all the providers out there to basically fill in the pieces, but I'm not sure they're going to have the capability of doing that," Linthicum says. [3:37 -- 8:02]
Will public cloud vendors embrace hybrid IT?
One size does not always fit all in cloud. Enterprises see the value in having a hybrid cloud model, where using both public and private cloud services gives them more control and increased security. But not all cloud providers have jumped on the hybrid or multicloud bandwagon.
"I don't think the hybridness requirement is completely appreciated at the big cloud providers," Banerjee says. "Ultimately, the economics of [Amazon Web Services] or Google cloud rest on sucking in all the compute and all the data and providing an economy of scale that is irresistible for lines of business. But there are going to be a number of reasons why the enterprise does not yield its entire data set and computation to one vendor, for all time."
Not all enterprises workloads will land in Google, Amazon Web Services (AWS) or Azure, Linthicum agrees. To better meet enterprise needs, some public cloud vendors will need to evolve their culture, executive team or mindset.
"Google, as well as the other cloud providers, need to do a better job in understanding that it's a complex sell … and, guess what? You have existing legacy systems. IBM mainframes are going to be around for years. You have to work and play well with those environments." [9:55 -- 13:00]
What else do providers need to attract enterprise customers?
To win in the enterprise, the big three public cloud vendors need to take other steps beyond hybrid cloud.
For example, after the development process is over, enterprises still need more from their providers."[They] still need the production environments to be highly efficient, they need isolation, they need integration with all the legacy and all the upstream and downstream apps," Banerjee says .
Meeting enterprise needs not only benefits the enterprise, but also the provider. "They're missing opportunities because they're not necessarily willing to work with these very complex environments -- AWS' unwillingness to work in hybrid cloud environments, for example. But that's got to change," Linthicum says. [13:00 -- 15:37]
Machine learning in hospitals
The benefits of machine learning are cropping up in a variety of industries -- including healthcare. Learn how Beth Israel Deaconess Medical Center in Boston joined forces in 2014 with MIT to improve the efficiency of the labor and delivery department through machine learning algorithms.
How does machine learning benefit enterprises in the cloud?
One of the new cloud technologies gaining traction now is machine learning. AWS, Azure and Google have all bulked up their big data services, helping to provide storage for the massive amounts of data used in machine learning algorithms. By combining cloud and machine learning, "we're going to build these really powerful, thinking systems that are basically going to have access to gobs and gobs of information, where they can make intelligent decisions and basically learn as they go," Linthicum says.
Machine learning in cloud can be a powerful tool for enterprises, helping them learn more about their customers or even their own operations. For example, by using a machine learning algorithm, a retail business can detect sales patterns and pull obsolete products off the shelf more quickly. Using such information can save money and allow businesses across industries to make better operational-level decisions.
"I think there is a set of ease-of-use improvements, as well as fundamental, qualitative changes, like self-driving cars, that are going to come out of this thing that we are calling machine learning … [and] they will touch all our lives in incredibly important small and big ways," Banerjee says. [19:20 -- 25:56]
Understand the hybrid cloud model
Compare the top providers' container services
Migrate to a hybrid cloud
Build a hybrid cloud management plan