Modern infrastructure and machine intelligence are two trends that will have an impact on every entity of the industry including platform companies, independent software vendors, global system integrators and emerging startups.
The convergence of modern infrastructure and machine intelligence will create a perfect storm for the IT ecosystem. While a few will take advantage of this trend, the majority will struggle to stay relevant in the industry. This trend will result in much bigger disruption than what was caused by cloud computing and SaaS.
What are modern infrastructure and machine intelligence? Are they same as cloud computing and artificial intelligence? How is machine intelligence different from machine learning and artificial intelligence?
Cloud computing has fundamentally changed the IT infrastructure landscape. The attributes like self-service, pay-by-use, elasticity and automation added a new dimension to infrastructure management.
Modern infrastructure is not the same as cloud computing. It is predominantly driven by emerging use cases, and business problems that cloud alone cannot adequately address. The rise of containers, container orchestration, microservices, cloud-native architectures, container-defined-storage and container-defined-networking lead us to the next phase of infrastructure.
Next-generation infrastructure will be built on the firm foundation laid by containers. Like the way VMs abstract the underlying physical hardware, containers will make virtual machines irrelevant. Container images will become the de facto mechanism for packaging, distributing and deploying software. With containers becoming the new VMs, Infrastructure as a Service (IaaS) will gradually transform into Container as a Service (CaaS).
Serverless computing and edge computing will become the layers that expose core computing capabilities. Applications will run in a distributed mode where certain components will be deployed as containers in the cloud and the remaining running in the edge layer. A significant part of the code will move into serverless computing to handle events triggered by the cloud and edge. Next generation content delivery networks (CDN) will be able to run code across edge locations making the applications truly distributed.
Infrastructure as Code, DevOps and automation are essential for managing modern infrastructure. They become as ubiquitous as Linux shell scripts for automating workload deployment, scaling and management.
In the era of AI, GPUs are becoming as prominent as CPUs. With container platforms gaining access to GPU clusters, data scientists can package algorithms as container images which are deployed at scale to perform training of the machine learning models. The trained models are also deployed as containers used for scalable inference.
One technology that is at the front and center of modern infrastructure is Kubernetes. It is winning mindshare and market share to become the Linux of containers. From enterprise data centers to web-scale public cloud environments, Kubernetes will form the core infrastructure layer for running next generation applications. The cloud-native ecosystem that is building the tools for Kubernetes will immensely benefit from this trend. This platform would be running the whole spectrum of workloads ranging from stateless web applications to line-of-business applications to large enterprise applications like ERP and CRM.
Kubernetes and related tools are all set to become the core building blocks of modern infrastructure.
Machine intelligence is the convergence of Internet of Things (IoT) and cognitive computing. The industry hasn’t realized the full potential of IoT. When machine learning and artificial intelligence are infused into IoT, we will experience the real power of intelligent devices.
Machine intelligence harnesses the power of the public cloud and edge computing. Powerful machine learning models will be evolved in the public cloud and deployed in the edge. From doorbells to CCTV cameras, AI will become an integral part of the devices. They will take advantage of AI running at the edge.
Today, AI and ML are confined to niche applications used only by a subset of the users. AI inference at the edge will put the power of machine learning in the hands of consumers. Machine intelligence combined with edge computing aims to democratize AI by making it accessible to the masses.
In the previous era of cloud, the focus was on the collaboration between developers and operations. With machine intelligence, the collaboration between data scientists and developers will become essential. Data engineers and data scientists will exploit public cloud for preparing the data and evolving accurate machine learning models, which are consumed by developers across desktop, mobile, web, and wearable environments.
Since PCs and servers are also devices that generate logs and metrics, machine intelligence will empower data center and cloud administrators with AIOps – an intersection of AI with operations where failures and downtimes are handled proactively.
Conversational AI will bring natural interaction between consumers and devices. Practically, every device and application will feature voice capabilities.
TensorFlow will be one of the key technologies that will drive machine intelligence. It is on its way to becoming the preferred framework running across the cloud and devices. Alternatives such as Microsoft CNTK, Apache MXNet, Caffe and Torch will also witness increased adoption.
The best way to ensure that companies stay relevant in the IT industry is to invest in modern infrastructure or machine intelligence. These are two trends that are set to become the next big thing of the industry.
The lethal combination of modern infrastructure and machine intelligence will drive the next phase of growth. Computing will be handled by containers, serverless and edge computing layers. Applications will be deployed in distributed environments running across the cloud, devices and the edge. They will consume AI capabilities irrespective of their deployment environments.
Kubernetes delivers the infrastructure layer while TensorFlow will be the intelligence layer powering the next generation of applications.