Alluxio founder also says pyTorch or TensorFlow will emerge as the top choice among machine learning frameworks.
Engineers who can manage structured and unstructured data will be the most in demand IT professionals in 2020 as companies build out their data technology infrastructure, according to Alluxio founder and CTO Haoyuan Li.
He describes 2020’s “super-power” skill as the ability to understand data as well as pick the right approach to analyze it.
Li also sees a bright year for hybrid cloud services in general and the expansion of cloud services in China as well.
Li’s company provides a data abstraction layer for computation frameworks, which allows applications to connect to several storage systems through a common interface.
Here are three of Li’s 2020 data predictions as well as reality checks from other industry experts.
China and the cloud
Li thinks that China will leapfrog from on-prem to massive cloud deployments for advanced workloads. He predicts that China will start using the cloud at a scale much larger than the US by adopting the public cloud for new use cases and for bursting in the cloud for peak loads.
SEE: Hybrid cloud: A guide for IT pros (free PDF)
Public cloud leaders in China such as Alibaba Cloud, Tencent Cloud and Huawei Enterprise Cloud will see dramatic growth that might outpace the growth of the cloud computing companies currently dominating the market.
At the “IDC FutureScapes: China ICT Market 2020 Predictions” event in December 2019, Kitty Fok, managing director of IDC China, predicts that by 2023, 60% of organizations will aggressively modernize legacy systems through extensive new technology platform investments, driven by escalating cyber threats and the need for new capabilities.
Thomas Zhou, associate vice president of IDC China, said that hybrid cloud deployment will support digital transformation on the IT platform level and that China’s digital economy will enter the fast lane in the next three years.
ICD predicts that by 2023, China’s digital economy will account for 51.3% of the country’s total economic output, with digital products and services comprising the lion’s share.
The year of hybrid cloud services
This is the year that all the predictions about hybrid cloud computing will become reality. Li reports that he is seeing large enterprises invest in the public cloud instead of adding capacity on-prem to Hadoop deployments.
He thinks companies will keep core enterprise data on-prem and compute will be burst to the cloud, particularly for peak demands and unpredictable workloads.
He also thinks that companies that provide technologies to achieve this will—like Alluxio—drive the rise of the hybrid cloud.
TechRepublic writer Jack Wallen also sees this as a 2020 trend with companies using a private cloud that can fail over to a public cloud as needed as the ideal way to leverage the cloud.
Dominant machine learning framework will emerge
Machine learning with models has reached a turning point, according to Li, and the next step is to operationalize model training efforts.
While there are several popular frameworks for model training, a leading technology hasn’t yet emerged. Just as Apache Spark is a frequent choice for data transformation jobs and Presto for interactive querying, 2020 will be the year a frontrunner will dominate the broader model training space with pyTorch or Tensorflow as leading contenders.
Richard Gall, co-editor at Packt Hub, also thinks that pyTorch and TensorfFlow are “going to be defining deep learning in the year – or years – to come.”
Gall writes that although TensorFlow has been leading the way for some time, pyTorch is gaining ground, thanks in part to the release of pyTorch 1.3 in October.
The full list of Li’s 2020 data predictions includes Kubernetes, Hadoop, and the merger of AI and analytics teams. Li founded Alluxio at UC Berkeley’s AMPLab and the company is backed by Andreessen Horowitz and Seven Seas Partners.