Cloud-native technology has undergone a really remarkable transformation. The foundation of cloud-native technology, Kubernetes, will celebrate its tenth anniversary in 2024. According to Industry Research Biz, it is expected to maintain its current position and grow at a Compound Annual Growth Rate (CAGR) of 18.51% during the next few years, reaching USD 5575.67 million by 2028.
Cloud Native continues to influence and redefine how organizations function and provide value to their clients by including both micro-trends and IT macro-trends.
The mature, developing, and changing cloud-native ecosystem is a topic of interest for us at Percona as we close up 2023 and consider what lies ahead in 2019.
Recap of KubeCon NA 2023
Matthew Kane, Content Marketing Director PerconaThere was no doubt about the subject of KubeCon NA: AI and Large Language Models (LLMs). Throughout the keynotes, the main topic was how Cloud Native and Kubernetes assist companies in embracing the AI future. And that makes sense, as Kubernetes is gradually becoming into the Platform, which is what it was meant to be.
The publishing of the CNCF platform whitepaper and the addition of a Platform Engineering day to the next KubeCon show that the topic of platform engineering has advanced significantly. Here at Percona, we see an increasing number of businesses using Kubernetes to provide services to their teams, accelerating software delivery and propelling expansion.
Adding orchestration on top of orchestration is possible in the community using Declarative GitOps management, ArgoCD, and Flux. We verified the CNCF GItOps Microsurvey data - 91% of developers and engineers we spoke with at the conference are already using GitOps.
In the Dynatrace Kubernetes in the Wild 2023 study, it is reported that a noteworthy 71% of respondents are presently using databases in Kubernetes (k8s), with a rise of 48% year over year. This result is consistent with what was seen during the Data on Kubernetes (DoK) day a year ago, when talks about this issue moved from specialized, tech-focused dialogues to a broader, enterprise-level interest in implementing a variety of use cases. Based on these factors, it seems that databases on K8s are still in their infancy and will probably continue to gain traction in the coming years.
Forecasts
Multi-Cloud is a Requirement
Although this tsunami has been accumulating for years, we anticipate its zenith in 2024. Hashicorp commissioned a 2023 Forrester study, and 61% of participants said they have upgraded, expanded, or launched a multi-cloud strategy. In 2024, we anticipate that number to increase.
Almost all of the vendors we interacted with at Kubecon and the vendors themselves have some kind of multi-cloud strategy or necessity. This may sometimes result from acquisitions or mergers that are necessary. A key component of contemporary infrastructure planning is often avoiding cloud vendor lock-in. It is now widely used, therefore if it isn't a part of your plan, you are lagging behind.
Adopting this method has several commercial benefits:
- Adaptability in taking use of cloud-vendor innovations to develop more quickly
- Freedom from vendor lock-in, which increases bargaining leverage
- Higher RPO and RTO for application and database architecture
- Complying with client security and governance criteria
Percona has taken this value into consideration while designing its Operators for PostgreSQL, MongoDB, and MySQL. Our goal is for users of our technology to be able to use any public or private cloud environment to install their vital open source databases and apps. To make managing and operating your clusters easier, the operator comes with all the database automation needed to maintain a highly available, robust, and secure database.
Make things Simple and Secure
It is evident from examining a number of State of Kubernetes reports (from VMware, RedHat, and SpectroCloud) that platform engineering teams are primarily concerned with complexity and security.
Simplifying may occur from several perspectives. While deployment is mostly resolved, management and operations remain unresolved. We anticipate seeing a range of fundamental fixes and tools to automate migrations, scaling, upgrades, debugging, and other tasks.
Operators play a crucial role in resolving the complexity issue by eliminating the need to understand internal application setup details and K8s primitives. Additionally, they free up labor and let engineers concentrate on developing applications rather than platform engineering tasks. There will be an influx of new operators as well as an evolution of current operators, who will provide capabilities that either match or surpass managed services available on public clouds.
According to the most recent analysis on Kubernetes adoption, security, and market trends in 2023, 67% of respondents said that worries about Kubernetes security have caused them to postpone or slow down implementation. Furthermore, 37% of respondents said that a container/Kubernetes security breach cost them money or lost them customers.
With the vulnerability of open source software ranking among the top concerns, supply chain attacks (such as the SolarWinds attack and vulnerabilities like Log4Shell and Spring4Shell) increasing in frequency, and container and Kubernetes strategies becoming more prevalent, cybersecurity and operational understanding are becoming increasingly important in development.
The increasing complexity of contemporary systems, particularly those running on platforms like Kubernetes, is a major security risk as well. This emphasizes the need for unified threat models and vulnerability detection tools. For teams and infrastructures to share similar knowledge and patterns, standardization and cooperation are essential. To increase overall security in cloud systems, repositories for memory-safe patterns should be created.
Most of the responders to RedHat's security research survey are working on a DevSecOps project. DevSecOps, a phrase that encompasses methods and technologies that enable security to be incorporated into the application development life cycle instead of being viewed as a distinct process, is being adopted by the majority of enterprises. Nevertheless, 17% of companies do not have any DevSecOps efforts and instead manage security independently from DevOps. As a result, they could pass on the advantages of including security into the SDLC, such increased software delivery quality, speed, and efficiency.
MLOps and AI
With an emphasis on making the development and deployment of AI applications, including model training, easier, Kubernetes has emerged as a new web server for several production AI workloads. With IBM and META at its helm, the recently established Open Source AI Alliance pledges to uphold open-source AI. It is made up of several groups from different industries, such as academic, public, nonprofit, hardware, software, and hardware. The objective is to conduct scalable and distributed training tasks for popular frameworks like PyTorch, TensorFlow, MPI, MXNet, PaddlePaddle, and XGBoost as well as jointly build tools and programs enabling open development.
Users are calling for AI to be more transparent and cooperative when incorporating AI and machine learning into cloud-native infrastructures. It is impossible to overlook the patterns that are emerging from ‘AI Advancements and Ethical Concerns.’
Transparent AI frameworks and ethical principles must be included throughout application development in order to address ethical problems and biases. Consumers will place a higher priority on AI education and efficiency in order to address moral and legal issues. This signals the end of a chaotic period and opens the door to faster innovation, more efficiency, and standard operating procedures.
In Summary
At Percona, we place a high value on keeping ahead of market trends by using the knowledge and skills of our staff and abiding by industry best practices.
We've always taken care to prioritize security while developing software, and integrating multi-cloud deployment into our offerings has been a key component of our approach. We take extra security measures since we are committed to open source software. These measures ensure operational security by using best practices and concepts like least privilege, security in layers, and role/responsibilities separation via software controls and policy. Furthermore, we continuously add new sharding features to our roadmaps with multi-cloud in mind. One example is the planned Shard-per-location support in the Percona Operator for MongoDB.
In addition, we're not holding back when it comes to introducing cutting-edge features to cater to any fresh, emerging trends and rocking the cloud-native community. Did you say: “More Simple Kubernetes”? After a year of intense effort, storage autoscaling for databases in Kubernetes is now here and will be released in Q1, 2024. Our Operators and Everest will be able to enjoy a serverless-like interface thanks to this completely automatic scaling and tweaking. The endpoint will be provided to developers without requiring them to think about resources or adjusting at all. It doesn't need for human involvement and is worry-free.
Lastly, our team decided to offer support for the pgvector extension to Percona-powered database software in order to add vector-handling capabilities. This move was motivated by the growing popularity of generative AI and engines like OpenAI or Bard.
We will keep pushing the envelope for our community and the rest of the globe, with a constant emphasis on innovation to speed up development for everybody.
The creation, modification, and removal of members in your Percona Distribution for MySQL, MongoDB, or PostgreSQL environment are automated by the Percona Kubernetes Operators.
