Mainstream breakthroughs for AI and machine learnings
In 2023, we are going to see increased awareness of the capabilities of AI and ML, including DALL-E-2, a new AI system that can create realistic images and art from a description in a natural language, and GPT-3, an autoregressive language model that uses deep learning to produce human-like text. We will only see better and more unbelievable progress in this space. We will also see how faster AI chips by companies like Qualcomm, Intel, Google and NVIDIA make AI more consumable than ever.
Additionally, running AI at the edge will be critical in the future to solve problems as closest to the data being captured/ingested, which cannot be done fast enough when pushed to the central cloud. In 2023 and beyond, we will see the adoption of AI chips embedded inside servers at the edge and related workloads.
Multiple cloud adoption and distributed / edge computing
In 2023, we will see more organizations using multiple clouds to run their workloads, supporting numerous vendors to run workloads where they are best served. Most organizations have just started adopting cloud migration and cloud-native due to the COVID-19 pandemic, with the demand to work from anywhere catalyzing the adoption and proving it works. We are also seeing edge computing and emerging solutions start taking shape, and distributed computing and edge computing are becoming real business opportunities at scale.
In the next year, we will see distributed edge cloud become the standard cloud perception and approach. As we trend toward distributed cloud and multi-cloud, zero-trust network access provides better security than ever before, so secure access service edge (SASE) uptake is expected.
Increased adoption of low-code will accelerate application execution speed
In 2023, we will see higher adoption of low-code to address new business demands and better time-to-market. During the coronavirus outbreak, a boom in remote software production fueled interest in low-code and no-code programming resources. This growth in low code has been influenced by the increased need for personalized tech applications to aid digitalization, which has spurred the advent of citizen developers outside of IT. Most low-code and no-code work focus on developing automation. IT leaders are stressed to significantly accelerate application execution speed and time-to-value due to digital market acceleration.
Increase in quantum computing resistant cryptography will improve data protection
In 2023, we will see increased adoption of quantum computing (QC) resistant cryptography. It is feared that recent QC advancements can seriously risk current encryption standards we have today, and that in the next 1-2 years QC will be powerful enough to break almost all security cryptography algorithms used today. It has already been reported that hackers are purposefully stealing highly encrypted data which they cannot decrypt today and storing it as they will be able to decrypt it in the next year and onward using QC.
We are seeing a call from the industry for organizations to start adopting Quantum Computing Resistant Crypto to better protect their data. This means we will see the adoption of new algorithms and ciphers to protect data in ways that cannot be broken by the QC power we have today and in the future.
Confidential computing for better security
In 2023, we will see further adoption of confidential computing as it will become the new minimum in years to come. Today, confidential computing is expensive as only specific CPU models from Intel and AMD support it, and Intel CPUs do not fully support all capabilities. This will change as more and more CPUs will natively support it. Confidential computing provides encryption of the entire server memory with a security key or each VM memory space with unique security keys. It ensures that memory cannot be read by any process, even by administrators or VM operators, as the system's memory is encrypted at the hardware level.
Confidential computing further secures public cloud shared environments and locks workloads to specific hardware and cloud, meaning even if someone steals a server or a VM, it cannot be used anywhere outside of the original place it was created, which is critical for sensitive data.
As confidential computing requires modern CPUs and chipsets to support this at the silicon level, the more common it is, the cheaper it will be. Intel currently supports memory encryption of the entire server memory but not per VM. This capability in their CPUs will change next year to be on par with AMD.