For Francois Paillier, the decision to take a hybrid cloud approach was made for him early in the pandemic, when his server literally went up in smoke. As the CEO and cofounder of the genomics startup CircaGene, he had immediately joined the race to genetically sequence the COVID virus, only to have his primary host catch fire. Improvising, his team fell back on their on-premises computer while racing to spin up new virtual machines. I had to pilot my business to go fully virtual, so it needed to be secure, scalable, and resilient, he said. And now we have a hybrid system that is much better, with different servers for different purposes.
Since the launch of Amazon Web Services 20 years ago, there has been a steady shift away from companies running their own hardware in favor of fully virtualized everything-as-a-service. But the persistence of legacy systems coupled with emerging technologies such as AI has CTOs seeking a platform-agnostic hybrid model that strives to have the best of both worlds. The goal isnt to slash IT costs but to give their companies the ability to deploy and iterate fasterat speeds no one thought possible until a few years agoand become true catalysts for innovation.
The original promise of the cloud was agility, said Talia Gershon, director of cloud infrastructure research at IBM, during a wide-ranging conversation at The Future of Hybrid Cloud, a virtual event hosted in partnership by Fast Company and IBM. How can we achieve business objectives faster? One way is to bring the cloud to new places while adding new capabilities. To explore what these applications and their business cases might look like, Gershon was joined by Paillier and Sam Carter, CEO of Moneo. Here are three key takeaways from their discussion.
In CircaGenes case, this meant learning how to sort tasks and stratify data by securing confidential genetic records both locally and in an encrypted cloud, while using another cloud for analytics and computation. The question of where do we go? is business-driven, Paillier said.
Moneo faced a similar challenge. As a cash-back rewards platform launched amidst the pandemic, its clientsconsumer packaged goods brands (CPGs)were desperate to harness its data for their suddenly all-important online advertising. Brands wanted to know who was buying their products, where they were buying them, and how they wanted to interact, Carter said. The pressure faced by CPGs and retailers was immense.
They, in turn, bombarded the startup with a constant stream of requests for real-time custom analyticsany scrap that would give them an edge as foot traffic to physical stores plunged. In addition to privately hosting its core operations, Moneo scrambled to create a replica of its data in IBMs public cloud to handle the mounting number of reports. They needed to work with someone who was quick, predictable, safe, and scalableand who could do all of those things immediately, Carter added.
Nimbly pivoting from a public- or private-only approach is one thing when youre a startup; its another altogether when your organization is nearly bankrupt from decades of technical debt. Imagine being the developers tasked with figuring out how to refactor millions of lines of code into microservices and bring them into a cloud-native model, Gerson said. Fortunately, AI can help analyze which pieces of the technology stack should be virtualized onto public clouds, which can be refactored efficiently, and which should be left alone. Thats where technologies such as Kubernetes and OpenShift are able to offer a consistent user experience while enabling you to run applications where they make sense, she added.
Those applications include such cutting-edge techniques as homomorphic encryption, which enables CircaGene to analyze private health records without needing to decrypt them. Its an approach perfectly suited to hybrid clouds, using a publicly accessible algorithm to run blind computation against data stored securely in private systems. But Pailliers original inspiration was personal: the plight of a friend with breast cancer whose reluctance to share her fears had delayed detection. I decided to develop a module that examines your encrypted DNA and returns the results with a private encryption key and a recommendation, he said. Above a certain level, you need to take action.
While less a matter of life or death, Moneo is also grappling with questions of data governance and consent. Countless consumers have found it easytoo easyto share personal data, with little recourse for opting out once they have opted in. Thats about to change with a wave of new global legislation and regulations aimed at protecting users privacy that has created a new headache for brands. Whos agreed to share what, and how does that change over time? Carter asked rhetorically. Weve become an attractive option for brands because we do something they currently cantand isnt really core to their business.
The timing of that statement is important. While Moneo, through its hybrid cloud reports, can now offer customers an innovative service, the same capability is inevitably tomorrows open-source commodity. In fact, it already isIBM Research has already launched a beta test of Fybrik, a cloud-based service for orchestrating secure data governance across companies and platforms.
This is a topic near and dear to our hearts in the research division, Gerson said. How can new technologies help your teams move faster while automating compliance and minimizing risk? Thats the vision guiding our technology roadmap for the future of cloud.
View original post here:
What's next: A look into the bright future of hybrid cloud - Fast Company