Cloudflare’s Cloud-Native Architecture: Delving into the Speed and Security of the AI Era
Dave Barnett, Head of SASE at Cloudflare, sat down with AI News to discuss the company’s cloud-native approach to delivering speed and security in the AI era.
Cloudflare’s Cloud-Native Approach
According to Barnett, Cloudflare’s cloud-native architecture allows the company to continually innovate in the digital space. Notably, a significant portion of their services are offered to consumers for free.
"We continuously reinvent, we’re very comfortable in the digital space. We’re very proud that the vast majority of our customers actually consume our services for free because it’s our way of giving back to society," said Barnett.
Focusing on AI
Barnett also revealed Cloudflare’s focus on AI during their anniversary week. The company aims to enable organizations to consume AI securely and make it accessible to everyone. Barnett says that Cloudflare achieves these goals in three key ways.
"One, as I mentioned, is operating AI inference engines within Cloudflare close to consumers’ eyeballs. The second area is securing the use of AI within the workplace, because, you know, AI has some incredibly positive impacts on people … but the problem is there are some data protection requirements around that," explains Barnett.
"Finally, is the question of, ‘Could AI be used by the bad guys against the good guys?’ and that’s an area that we’re continuing to explore," he added.
Zero-Trust Models
Regarding security models, Barnett discussed the evolution of the zero-trust concept, emphasizing its practical applications in enhancing both usability and security. Cloudflare’s own journey with zero-trust began with a focus on usability, leading to the development of its own zero-trust network access products.
"We have servers everywhere and engineers everywhere that need to reboot those servers. In 2015, that involved VPNs and two-factor authentication… so we built our own zero-trust network access product for our own use that meant the user experiences for engineers rebooting servers in far-flung places was a lot better," says Barnett.
"After 2015, the world started to realise that this approach had great security benefits so we developed that product and launched it in 2018 as Cloudflare Access," he added.
Leveraging NVIDIA GPUs
Cloudflare’s innovative strides also include leveraging NVIDIA GPUs to accelerate machine learning AI tasks on an edge network. This technology enables organizations to run inference tasks – such as image recognition – close to end-users, ensuring low latency and optimal performance.
"We launched Workers AI, which means that organisations around the world – in fact, individuals as well – can run their inference tasks at a very close place to where the consumers of that inference are," explains Barnett.
"You could ask a question, ‘Cat or not cat?’ to a trained cat detection engine very close to the people that need it. We’re doing that in a way that makes it easily accessible to organisations looking to use AI to benefit their business," he added.
Supporting Machine Learning
For developers interested in AI, Barnett outlined Cloudflare’s role in supporting the deployment of machine learning models. While machine learning training is typically conducted outside Cloudflare, the company excels in providing low-latency inference engines that are essential for real-time applications like image recognition.
Conclusion
Our conversation with Barnett shed light on Cloudflare’s commitment to cloud-native architecture, AI accessibility, and cybersecurity. As the industry continues to advance, Cloudflare remains at the forefront of delivering speed and security in the AI era.
FAQs
- What is Cloudflare’s approach to AI?
Cloudflare focuses on operating AI inference engines within their platform, securing the use of AI within the workplace, and exploring the potential misuse of AI. - What is Cloudflare’s zero-trust model?
Cloudflare’s zero-trust model involves operating AI inference engines within their platform, securing the use of AI within the workplace, and exploring the potential misuse of AI. - What is Cloudflare’s role in supporting machine learning?
Cloudflare provides low-latency inference engines that are essential for real-time applications like image recognition, making it easily accessible to organisations looking to use AI to benefit their business.
