Use of the cloud has grown significantly. Cloud computing has become so popular, many startups are taking a “cloud-first” approach (the deliberate practice of considering use of the cloud first). The cloud offers tools, data, and scalable computing power that would be difficult to access in other ways. Cloud architecture, whether it be public, private, or hybrid, is the most cost effective and efficient means of scaling, both horizontally and vertically. In a recent cloud computing survey, done by the venture firm North Bridge, it was shown 50 percent of the organizations surveyed used a cloud-first philosophy, or used the cloud exclusively for their marketing needs.
The enormous growth of cloud use has caused analyst firms to continuously correct their forecasts. For instance, Forrester recently predicted the public cloud market would, by 2020, grow to $236 billion. However, their previous projections for 2020 were significantly lower, forecasting in 2014 that 2020’s cloud market growth would only be $191 billion, and in 2011 their growth prediction for 2020 was $159 billion. It would seem the cloud’s popularity has been as difficult to predict as its evolution.
Containers and the Cloud
Containers, in a nutshell, break apps into smaller, self-contained and easily managed packages of code. This code is also bundled with the software apps needed to function independently of the host server. Containers offer a standardized way to include an application’s code, its configurations, and its dependencies into one single package. Containers allow work and apps to be transferred almost seamlessly from a private cloud to a public cloud, and back again. According to ESG research, containers made up about 19 percent of the hybrid cloud workloads in 2018, and it predicts that by 2020, containers will be used 33 percent of the time in hybrid cloud workloads.
The Docker container has become quite popular. From the start, Docker has been open sourced, which helped its popularity significantly in the technology market. Its popularity, in part, is also based on its use as a replacement for virtual machines, and because it coincided with timing of the DevOps revolution. Additionally, Docker containers resolve software delivery and deployment issues that have plagued cloud-using organizations.
Kubernetes is a popular orchestration system used for Docker containers. It is more comprehensive than the original Docker Swarm, and is used to coordinate node clusters at scale in an efficient manner during production. Kubernetes (occasionally shortened to “kube”) is an open-sourced system, was developed by Google originally, and organizes containers into coordinated, logical units designed for transfer and use within the cloud.
Both Docker and Kubernetes will enjoy great popularity in 2020.
Open Source Software and the Cloud
Open source software has become extremely popular with many organizations, especially startups. They commonly integrate open source software into their operations, and may build their entire
infrastructure around it. The open source philosophy is thriving in the cloud computing ecosystem, with people realizing that sharing offers more benefits than hoarding (see synergy).
A recent survey by Black Duck Software showed that up to 60 percent of decision-makers in the industry are currently using open source software, and more than half are making contributions to open source projects. As organizations migrate their workflow to the cloud, open source software will play a key role in innovation in 2020 and beyond.
Interest in Serverless Computing Drops
Serverless Computing is a form of cloud computing that has the cloud provider running the server, and managing the allocation of resources. Pricing is determined by the actual amount of the resources used by an application instead of pre-purchased units based on capacity.
According to a recent survey by Cloud Foundry, serverless computing is losing popularity. After a sharp increase in the use of serverless computing in 2018, their 2019 survey showed a drop. In their last year’s survey, 19 percent used serverless computing, but that number has dropped to 15 percent in 2019.
A recent study from the University of California at Berkeley may explain the drop in interest. The Berkeley team, led by Eric Jonas, reported:
“Storage in serverless settings is one area where more work needs to be done. The stateless nature of serverless platforms makes it difficult to support applications that have fine-grained state sharing needs. This is mostly due to the limitations of existing storage services offered by cloud providers.”
They also suggested security and performance issues were concerns.
Private and On-Premise Clouds
A “private” cloud is defined as a form of cloud computing that is used by only one organization, or that ensures that an organization is completely isolated from others. Private clouds are typically leased on a monthly basis from a contractor. On-premises clouds are also private clouds, but owned by the organization, and often located in the company’s data center, rather than running remotely on rented host servers or paying for time in a public cloud.
Because private and on-premise clouds are dedicated to one single organization, their architecture can be designed to meet that organization’s specific needs. According to a report by LogicMonitor, it is predicted that in 2020, 41 percent of data workloads will take place in the public cloud, with 59 percent running on private or on-premise clouds.
A private or on-premise cloud offers flexibility, control benefits, and (with the right tools) greater security. These advantages are especially valuable for organizations with customization requirements, predictable workloads, and businesses in regulated industries. Also, private clouds are much more cost effective for projects requiring time-consuming experimentation.
A typical multi-cloud environment is usually made up of at least two public clouds and one private cloud, which is either operated by the business or managed by a third party. However, for start-ups or other cloudless organizations, it’s not necessary to start off with a private cloud. Public clouds can handle short term needs without support from a private cloud. The multi-cloud environment offers several benefits for organizations, including avoiding dependence on one cloud vendor—which can become both clumsily expensive and unnecessarily restrictive.
Currently, the competition between cloud vendors is fierce, resulting in a broad selection of tools at reasonable prices (but keep an eye out for expensive “specialty” tools or services). A business can maximize their cloud research results (and minimize their expenses) by flexibly using their multi-cloud resources. This philosophy requires staff who are comfortable and familiar with the clouds being used (wasting time on a public cloud costs money, though many clouds offer a very reasonable free trial period, allowing staff to learn their system).
Hybrid clouds (which do require a private cloud) are being adopted by many organizations that want to experience the benefits of using both a private cloud and public clouds. A hybrid cloud is linked to the public cloud, and allows organizations to switch back and forth between their own tools, and the tools offered by various cloud providers. This improves both speed and flexibility. A hybrid cloud environment can provide on-demand, externally-provisioned scalability. It can also provide emergency backup services and disaster recovery scenarios.
Organizations using a hybrid strategy grew to 58 percent in 2019 (up from 51 percent in 2018). One IDC survey taken from 2018 found 80 percent of the respondents had moved cloud workloads to a private cloud within the past year.
The survey also revealed they were planning to migrate another 50 percent of their public cloud workloads to a private cloud. This is happening because some clouds have decided to charge more for certain services, and it can be more cost effective to create an in-house cloud than to pay high-priced fees for regularly-used services and tools. The hybrid strategy will continue to grow in 2020.
Edge Computing and the Cloud
Edge computing uses the IoT (Internet of Things). It is a concept that offers greater efficiency by having IoT devices screen out unnecessary data, and send only the useful information to the cloud, rather than raw streams of it. IoT requires collecting and processing large amounts of data during real- time and with a low latency level.
Edge computing has become the technology used to accomplish this. A recent a study done by Grand View Research has predicted edge computing will gain $3.24 billion of the global market. Various industries, such as automotive, oil and gas, the power utilities, and aerospace, benefit from edge computing.
Artificial Intelligence and the Cloud
Among the businesses adopting AI technology, 70 percent are obtaining AI capabilities by way of cloud-based enterprise software. Enterprise software that is integrated with AI is the most popular way of acquiring AI capabilities (and the easiest). This software is typically cloud-based. Deloitte Global estimates that by 2020, enterprise software with cloud-based AI platforms will reach 83 percent among organizations using AI software. The cloud will continue to push for more AI implementations.
Changes in Cloud Architecture
In 2020, the architecture of public clouds will adjust to meet the growing demands of their clients, and many private clouds will be transformed into hybrid clouds, allowing them to link and interact with public clouds.
Public clouds will focus on the multi-cloud philosophy, expanding their services and tools, with the long-term goal of providing all the services their clients need and want. They will also make it easier for their clients’ clouds to form hybrid relationships with their clouds.
Multi-cloud architecture should provide an environment where organizations can create a secure and safe cloud environment that exists outside of the traditional infrastructure. The primary reason for using a multi-cloud strategy is to select the best cloud to meet specific goals, such as using one cloud to develop an app, and another cloud to research big data.
While shifting back and forth between cloud providers can be complicated, especially in the early stages, providers are currently working to make shifting back and forth between clouds easier and more efficient. As multi-cloud platforms become more efficient, the multi-cloud will continue to evolve.
Hybrid cloud architecture should allow for the quick and easy transfer of workloads to a public cloud. Public clouds are working to provide maximum agility for clients (including automated workload transfers).
Image used under license from Shutterstock.com