A look at how tech trends are converging in distributed cloud architecture.
With the shift of workforces online in 2020, Chief Technology Officers pivoted IT priorities to shoring up availability, back-up and disaster recovery capabilities. In general, the pace of digital transformation increased during the pandemic. An intention to take advantage of that momentum, positioning teams for achieving post-pandemic business objectives, is reflected in the key trends for 2021.
Those trends can be extrapolated from a recent CTO roundtable with these panelists:
- Andrew Trossman: Vice President of Innovation at the Royal Bank of Canada (RBC)
- Paritosh Bajpay: Vice President of Business Products from Verizon
- Rana el Kaliouby: co-founder and CEO of Affectiva
- Jason McGee: IBM Fellow and Vice President and CTO of the IBM Cloud Platform
Trend 1: Operationalizing artificial intelligence (AI)
As CEO Rana el Kaliouby explains, Affectiva is an artificial intelligence (AI) company whose mission is to humanize technology. Affectiva innovates its product to more accurately track the complex signals in human perception. Machine learning algorithms trained on great volumes of data, along with cloud computing resources necessary for iterative work at scale, are essential to what Affectiva offers clients whose business depends on understanding the subtleties of what customers and consumers want.
For many large-scale IT organizations, using AI cloud services to innovate app user experiences, improve and optimize operations (AIOps) and gain insights from all types of data remain parallel goals in an ongoing journey of adopting cloud native tools, practices and services.
While AI-related goals have been carried forward as next-year priorities in each of the past five years, other cloud adoption challenges preempt achieving them. The elephant in the room that is demanding attention at the beginning of 2021 is a security breach of US Government and Fortune 500 Corporation networks that is already the biggest in decades, but whose extent really has not yet been determined.
At this point, because most enterprise companies use more than one cloud provider for different purposes, the big challenge standing in the way of AI-infused operations is the distributed location of the data on which machine learning work might be done.
Should a company — to realize the ambition of giving data scientists easy access to data — build a lake where all sources of data converge across different clouds? Just the data egress and ingress costs for such an undertaking may currently be cost-prohibitive. And even if the costs are acceptable, how does a rapidly evolving business pick the right long-term site to build the lake?
An alternative to solving that problem is bringing the needed data preparation, machine learning and other artificial intelligence tools directly to wherever the data resides, which will likely be in multiple locations.
With the IBM Cloud Pak® for Data as a Service, for example, data science teams get everything they need to process data in place, without the burden of maintaining the IBM Cloud Pak software, and with the great benefit of being able to monitor and manage the relevant data operations from a single, consolidated view within IBM Cloud. IBM Cloud Pak for Data as a Service integrates with IBM Cloud Satellite, a distributed cloud as service offering.
In the case of Affectiva, they used IBM Cloud bare metal computing resources and advanced vision algorithms to stand up their human perception analytics service. To do that, Rana el Kaliouby and her teams brought data into the public cloud. With distributed cloud computing, however, as their client base grows, Affectiva will gain efficiency from being able to do the necessary model training and machine learning wherever the client has their data.
Trend 2: Evolving the foundation of cloud security
As an innovation officer at Royal Bank of Canada (RBC), Andrew Trossman needs to make public cloud services readily available to banks’ development and operations teams. The only way a financial institution like RBC can do that would be either by bringing the cloud services into on-premises environments where their business-critical workloads run or by moving those workloads onto a public cloud platform with trustworthy security and regulatory compliance safeguards.
