We are entering “The Data Era,” where the capability to collect, analyze, and secure data becomes increasingly crucial to an organization’s success. These capabilities are especially important for cloud service providers, who will power digital transformation in the same way electricity powered the industrial revolution more than 200 years ago. Staggering amounts of data is now created, transformed, stored, and transmitted in the cloud, accelerating the need for workload intelligence, data security, and user privacy.
At Microsoft Build conference this week, Microsoft announced new and expanded collaborations with Intel to help developers harness accelerated artificial intelligence (AI) in the cloud through Azure Machine Learning Hardware Accelerated Models powered by Project Brainwave*, and provide more secure, hardware-protected services with the public preview of Azure* Confidential Cloud. Both are great examples of the innovative, differentiated services that can be created with Intel® technologies and strong collaboration.
“Real-time AI” Ready for Developers
At last year’s Hot Chips conference, Microsoft unveiled Project Brainwave, a new deep learning acceleration platform leveraging Intel FPGAs, capable of handling extremely challenging deep learning models at “unprecedented” levels of performance and flexibility. In Azure, the FPGA-fueled architecture is economical and power-efficient, with a very high throughput. Microsoft stated that they can run ResNet 50, an industry-standard DNN requiring almost 8 billion calculations, without batching. This is a breakthrough in “real-time AI.”
Microsoft uses AI to enhance the precision and responsiveness of its own applications, such as Bing* Intelligent Search. At Build, Microsoft took another step forward and made AI acceleration available for its Azure cloud customers to run their own AI models on Azure accelerated hardware. The public preview of Azure Machine Learning models powered by Project Brainwave simplifies the journey of organizations to make smarter decisions, faster and at lower cost to customers.
Confidential Computing to Protect Data in Use
As cloud computing expands to a wider range of industries, it is imperative that cloud service providers address data privacy and protection head-on, and offer greater assurance to customers and users. Highly advanced threats are the “new normal” for the industry. Intel is committed to helping secure the digital world, and takes a collaborative approach with our customers and the security community to defend against modern and evolving threats.
Through years of joint innovation and R&D, Microsoft and Intel’s effort to enhance data protection in the cloud came to fruition in the Azure Confidential Computing announcement last August. Azure CTO, Mark Russinovich, introduced this breakthrough technology and the early access program to address one of the most pressing issues in the industry—protect data while it is in use.
Azure Confidential Computing enjoys hardware-based protection using Intel® Software Guard Extensions (Intel® SGX), which ensures that data is in a secure enclave or Trusted Execution Environment (TEE) when processed unencrypted. In this manner, Intel SGX extends Azure’s capability to protect data in use, in addition to data encryption at rest and in transit.
This week at the Build conference, Microsoft announced the preview of a new confidential compute virtual machine series for Azure enabled with Intel SGX technology as part of their broader Azure confidential computing efforts. As part of this offering, customers can develop, test, and run highly sensitive applications with the Intel® SGX SDK to protect data confidentiality and integrity in use. To get started, sign up to request access to the preview.
To find out more about Intel’s joint innovations with our cloud service providers, go to http://www.intel.com/csp.