In five years, machine traffic will surpass human traffic by a thousand times
- Machine traffic will surpass human traffic by a thousand times in 5 years
- The current growth of AI bots and IoT is accelerating the load on networks
- Cloudflare offers intelligent management and edge computing
- The optimization of AI accelerators will help reduce overload
At the financial briefing following the release of the Cloudflare report, Thomas Seifert stated that in the next five years, traffic generated by non-humans will surpass human traffic by a thousand times. According to him, this is not related to a sharp decline in user activity, but reflects the rapid growth of automated requests from AI agents, bots, and Internet of Things devices.
According to the CFO, machine traffic already constitutes a significant portion of network load today, but its growth rate is accelerating two to three times compared to traditional user activity. Users continue to use the internet for communication, shopping, and paying for services, but their share in the overall traffic picture will increasingly be 'shadowed' by software requests.
Why the forecast is important
Among the factors contributing to such a surge are the widespread adoption of generative AI models, automatic monitoring systems, and the increase in the number of connected devices. These technologies require constant requests to servers, often in real-time, leading to an increase in the volume of transmitted data.
Zaifert noted that in the past his forecasts sometimes did not come true, however, current trends are supported by statistics on the use of cloud services. If no measures are taken, the internet may become less convenient for real people, and the infrastructure may become overloaded.
Response options
Cloudflare offers several directions to mitigate the effect: intelligent traffic management, implementation of more efficient protocols, acceleration of data processing using specialized AI accelerators and graphics processors, as well as expansion of edge computing. According to their estimates, modern AI accelerators are not operating at full capacity, which opens up opportunities for optimization.
Industry experts believe that such measures will help maintain a balance between the needs of people and the growing load of automated systems. In addition, they help reduce latency and increase the resilience of networks to peak loads.
Overall, Zeifert's prediction emphasizes the need to rethink the architecture of the internet: from centralized data centers to more distributed solutions, where request processing occurs closer to the end user. This could become a key factor in maintaining service availability amid the rapid growth of machine traffic.
Source: 3DNews
