5 IT infrastructure trends: forecast for 2019
We decided to talk about the directions of cloud infrastructure development that IT experts and major technology publications consider the most promising in 2019. Under the cut - about serverless computing, "personal" quantum computers and AI systems.
/ photo 16: 9clue CC BY
Serverless architecture (serverless) is an approach in which third-party providers take on the task of maintaining servers. Thus, developers and system administrators do not have to spend resources on iron maintenance.
Serverless systems offer all the necessary functionality in the form of services. They can be either individual solutions (like the Fn Project ) or entire ecosystems (for example, Apache OpenWhisk or proprietary Azure).
Serverless architecture, according to Gartner, will be one of the hottest trends this year. Now less than 5% of companies worldwide use the serverless approach, and by 2020 this figure will increase to 20%. Already, there is a special literature , frameworks, conferences are held ( including in Russia ).
One of the drivers of the growing popularity of serverless systems is to reduce the cost of configuration and iron content. It is believed that system administrators spend about 30-40% of their time on setting up equipment. Serverless architectures release these resources - the efforts of engineers can be directed to the development of the company's products and services.
It is also expected that the “boom” of machine learning algorithms will affect the popularity of the serverless approach . Cloud vendors are making a lot of effort to ensure the availability and scalability of IT infrastructure. And these are the two most important components in the development of intelligent systems.
According to a study by Rightscale from 2018, about 80% of companies (out of a thousand respondents) have already switched to multi-cloud IT infrastructure. This year this figure will grow - in Gartner they say that by 2020 it will reach 90%.
The reasons for the popularity of multi-cloud are the cost savings and flexibility of using the infrastructure. Companies can choose the services of several cloud providers whose services suit them at a price. At the same time to simplify the migration of applications between different environments are increasingly used containers. According to the company ESG Research, by 2020, 33% of all workloads in the cloud will be containerized (today this number is 19%).
Analysts call one of the most popular containerization technologies Kubernetes. It is expected that in 2019 more solutions will appear on the market to simplify the implementation of this orchestration tool. For example, at the end of last year, VMware acquired Heptio, a startup developing solutions for deploying Kubernetes. Kubernetes is also developing IBM as an instrument for creating hybrid cloud solutions, which acquired Red Hat for this purpose (also in 2018).
According to forecasts of Deloitte, in 2019 will be the first commercial-5G network. Already by 2020 the number of users of new generation networks will be 15–20 million people. 5G plan to use to support the work of IoT-devices of industrial systems and " smart cities ". In this case, the number of gadgets of the Internet of things in the coming years will grow exponentially .
To cope with this load, telecommunications companies must ensure a high level of infrastructure scalability. Here cloud technologies will come to their aid .
/ photo Derek Finch CC BY (cropped image)
VMware developsVirtual solutions that simplify work with the core of the wireless network architecture and offer customers new features. An example would be VMware vCloud NFV for working with services and applications in a private and public cloud environment. Nokia created a similar platform specifically for operators - the service uses OpenStack technology to manage containers and virtual machines.
In this case, not only the cloud will help the spread of 5G, but, conversely, the network of the new generation will make cloud platforms more popular. In Deloitte they say that 5G will allow residents of remote regions where there is no high-speed Internet to use clouds.
IT companies are developing ever more powerful quantum processors. Last year, Intel introduced the 49-qubit chip, and Google - the 72-qubit chip .
Some devices already on the basis of qubits can be accessed via cloud services, such as Alibaba Cloud and the IBM the Q . The quantum machines of these IT giants help research organizations and universities to create complex models that are difficult to give to “classic” computers, for example, to model the behavior of a person’s immune system.
/ photo IBM Research CC BY
At the same time, 2019 could be the year of “personal” quantum machines . In early January, IBM at CES 2019showed Q System One - a prototype of a commercial quantum computer for business. This is a 20-qubit machine that, unlike other quantum computers, can operate outside the lab. The necessary conditions for the functioning of the computer are created in a special sealed container-case.
There is also a chance that in 2019 a “quantum superiority” will finally be achieved. This is the moment in time when machines with qubits begin to solve a number of computational problems faster than classic computers. Such tasks, for example, include the enumeration of keys SHA-256. To prove the quantum superiority hypothesis, at the beginning of the year, Google plans to hold a performance competition with the NASA supercomputer. The results of the experiment will be known in the summer.
Machine learning will be used more frequently in managing multi-cloud systems and equipment in the data center. Such technologies are called AIOps - AI for IT Operations. The rise in popularity of AIOps is predicted by Network World and Gartner. In 2019, 25% of organizations will use intelligent algorithms to manage their IT infrastructure.
AI systems for working with the cloud are already available in Oracle and Microsoft services . They help predict possible system failures and optimize workloads. There is a case where the Azure Cloud Learning Algorithm predictedFailure of one of the disks 20 minutes before it occurred. This allowed to quickly migrate the load from the failed server.
Platforms for managing equipment in the data center are also developing - this service was created by HPE. According to company representatives, the platform helped customers reduce operating costs by 79%, and the number of server failures - by 86%.
What we write about in the First blog about corporate IaaS:
Posts from our Telegram channel:
/ photo 16: 9clue CC BY
1. Demand for serverless computing will grow
Serverless architecture (serverless) is an approach in which third-party providers take on the task of maintaining servers. Thus, developers and system administrators do not have to spend resources on iron maintenance.
Serverless systems offer all the necessary functionality in the form of services. They can be either individual solutions (like the Fn Project ) or entire ecosystems (for example, Apache OpenWhisk or proprietary Azure).
Serverless architecture, according to Gartner, will be one of the hottest trends this year. Now less than 5% of companies worldwide use the serverless approach, and by 2020 this figure will increase to 20%. Already, there is a special literature , frameworks, conferences are held ( including in Russia ).
One of the drivers of the growing popularity of serverless systems is to reduce the cost of configuration and iron content. It is believed that system administrators spend about 30-40% of their time on setting up equipment. Serverless architectures release these resources - the efforts of engineers can be directed to the development of the company's products and services.
It is also expected that the “boom” of machine learning algorithms will affect the popularity of the serverless approach . Cloud vendors are making a lot of effort to ensure the availability and scalability of IT infrastructure. And these are the two most important components in the development of intelligent systems.
2. More companies will start using multi-cloud
According to a study by Rightscale from 2018, about 80% of companies (out of a thousand respondents) have already switched to multi-cloud IT infrastructure. This year this figure will grow - in Gartner they say that by 2020 it will reach 90%.
The reasons for the popularity of multi-cloud are the cost savings and flexibility of using the infrastructure. Companies can choose the services of several cloud providers whose services suit them at a price. At the same time to simplify the migration of applications between different environments are increasingly used containers. According to the company ESG Research, by 2020, 33% of all workloads in the cloud will be containerized (today this number is 19%).
Analysts call one of the most popular containerization technologies Kubernetes. It is expected that in 2019 more solutions will appear on the market to simplify the implementation of this orchestration tool. For example, at the end of last year, VMware acquired Heptio, a startup developing solutions for deploying Kubernetes. Kubernetes is also developing IBM as an instrument for creating hybrid cloud solutions, which acquired Red Hat for this purpose (also in 2018).
3. Clouds will help telecom operators with 5G sweep.
According to forecasts of Deloitte, in 2019 will be the first commercial-5G network. Already by 2020 the number of users of new generation networks will be 15–20 million people. 5G plan to use to support the work of IoT-devices of industrial systems and " smart cities ". In this case, the number of gadgets of the Internet of things in the coming years will grow exponentially .
To cope with this load, telecommunications companies must ensure a high level of infrastructure scalability. Here cloud technologies will come to their aid .
/ photo Derek Finch CC BY (cropped image)
VMware developsVirtual solutions that simplify work with the core of the wireless network architecture and offer customers new features. An example would be VMware vCloud NFV for working with services and applications in a private and public cloud environment. Nokia created a similar platform specifically for operators - the service uses OpenStack technology to manage containers and virtual machines.
In this case, not only the cloud will help the spread of 5G, but, conversely, the network of the new generation will make cloud platforms more popular. In Deloitte they say that 5G will allow residents of remote regions where there is no high-speed Internet to use clouds.
4. Quantum machines will become more accessible to users.
IT companies are developing ever more powerful quantum processors. Last year, Intel introduced the 49-qubit chip, and Google - the 72-qubit chip .
Some devices already on the basis of qubits can be accessed via cloud services, such as Alibaba Cloud and the IBM the Q . The quantum machines of these IT giants help research organizations and universities to create complex models that are difficult to give to “classic” computers, for example, to model the behavior of a person’s immune system.
/ photo IBM Research CC BY
At the same time, 2019 could be the year of “personal” quantum machines . In early January, IBM at CES 2019showed Q System One - a prototype of a commercial quantum computer for business. This is a 20-qubit machine that, unlike other quantum computers, can operate outside the lab. The necessary conditions for the functioning of the computer are created in a special sealed container-case.
There is also a chance that in 2019 a “quantum superiority” will finally be achieved. This is the moment in time when machines with qubits begin to solve a number of computational problems faster than classic computers. Such tasks, for example, include the enumeration of keys SHA-256. To prove the quantum superiority hypothesis, at the beginning of the year, Google plans to hold a performance competition with the NASA supercomputer. The results of the experiment will be known in the summer.
5. Intelligent systems will manage the cloud
Machine learning will be used more frequently in managing multi-cloud systems and equipment in the data center. Such technologies are called AIOps - AI for IT Operations. The rise in popularity of AIOps is predicted by Network World and Gartner. In 2019, 25% of organizations will use intelligent algorithms to manage their IT infrastructure.
AI systems for working with the cloud are already available in Oracle and Microsoft services . They help predict possible system failures and optimize workloads. There is a case where the Azure Cloud Learning Algorithm predictedFailure of one of the disks 20 minutes before it occurred. This allowed to quickly migrate the load from the failed server.
Platforms for managing equipment in the data center are also developing - this service was created by HPE. According to company representatives, the platform helped customers reduce operating costs by 79%, and the number of server failures - by 86%.
What we write about in the First blog about corporate IaaS:
- Why corporate customers use VMs, not containers
- How IaaS helps grow your business: three tasks that the cloud will solve
- Serverless computing in the cloud - a trend of modernity or necessity?
Posts from our Telegram channel: