10 predictions about data centers and cloud computing in 2019
1. Edge computing has developed completely but needs a business model
This is not difficult to find out. Everyone likes the idea of edge computers. Data center operators see this as an opportunity to reduce the load for central servers and businesses see this as an opportunity to have response times below 10 milliseconds. Vendors like Vapor IO and Schneider Electric are coming out with different models to place at base stations and 5G is starting to deploy.
The problem is, who pays for it? That has not been resolved. Will it fall into mobile providers, or will it fall into car manufacturers who want to connect cars? This industry has a long record of technology dreams and thinking about later business models, and edge computers are an expensive idea in an owner's search. That should be arranged in 2019.
2. The development of water radiator model
When Google launched version 3.0 of the new Processor AI chip, it also revealed that the company had switched to water cooling because the air was no longer enough. With the CPU reaching more than 200 watts and the GPU reaching 300 watts, air cooling simply meaning not cutting more. Water is thousands of times more effective in removing heat than air, and many companies are overcoming their fear of leaking coolant. In addition, in some cases, they have no choice. The demand for more processing power is pushing the transition to water cooling more than anything.
3. Apply AI to fix human errors
Data centers have thousands and thousands of moving parts: individual servers, cooling systems, electrical systems and networks to connect all. So far, they have been configured manually and once in place. But there is a new artificial intelligence (AI) layer, as shown by a startup called Concertio, that takes the AI in charge of optimizing the device through continuous monitoring and adjustment. There are in fact many cases in which AI is used as a constant, tireless screen for adjusting systems. Surely we will see more efforts in the future.
4. People will continue to develop data centers
Let's be honest, data centers are not dead. There are more computing needs than ever, especially when the advent of AI and the cloud has been proven to have costly disadvantages. However, that means data centers are being reused. Some workloads will be transferred to public cloud vendors, while others are assigned to data centers. This includes everything with huge data sets, such as BI, analytics and AI / ML, because moving them to the cloud is very expensive. Data centers are changing, becoming more flexible and powerful.
5. Workload moves from endpoint to data center
The data itself is worthless unless it is processed, and a smartphone is really not a device for it. Smart phones, tablets and PCs are heavy data collection tools but not suitable for analysis or any kind of AI, so data will be sent to the cloud for processing. The same thing applies to Internet of Things (IoT). 2019 car models will not process data, it will be sent to a data center for processing.
6. The rise of microservice and serverless computer
Virtualization is good, but it's heavy in terms of resources. It requires a full version of the operating system and that can limit the number of VMs on a server, even with a lot of memory. The solution is container/microservice and serverless computer. A container is less than 10 MB in size – too little compared to a few GB of memory for a full virtual machine with no server, where you run a single, even smaller, functional application.
7. AWS and Google focus on hybrid cloud
Amazon Web Services (AWS) and Google Cloud Platform participate in the cloud market without legacy and sales of a pure cloud game. Microsoft and IBM, on the other hand, have a huge legacy software already. Hybrid cloud therefore will be a wise choice for these tech giants to develop in this industry.
8. It will be a decisive year for Oracle
Oracle really needs to make some tough decisions this year - and quickly. Business operations on its cloud platform are leveling off and not keeping up with the four big guys (AWS, Microsoft, Google, IBM). Its license is still too complicated. It tried to win a major project called with the Pentagon Jedi from AWS and lost. It was quite silent about its hardware business, and now it has lost its cloud-based business leader. Oracle took a leap to the cloud gracefully like Microsoft, but if now the company does not have any improve, it will fall further.
9. Bare metal continues to grow
Bare metal means no software. You hire CPU, memory and storage. Then you provide your own software stack. So far, IBM is the largest proponent of bare metal storage, followed by Oracle. Bare metal is ideal for what Vikings call elevators and teleports, where you take your computing environment from the data center to the cloud provider.
IBM and Oracle are two major enterprise software vendors, so they will only want customers to continue using their software but run it in their data centers instead of switching to SaaS suppliers. However, AWS is entering the bare metal context, as well as some other large cloud vendors such as Internap, Equinix and Rackspace.
10. Cloud vendors fight for desktop computers
Microsoft is the only vendor viewing desktop computers as a means of connecting to the cloud, all other major cloud providers are interested in the virtual desktop market. And because Windows 7 will end in January 2020, 2019 will be a year of transition. The question is, will people jump to Windows 10, or will they embrace things like AWS WorkSpaces or Google Chromebook?
By: Emma Chavez