Uber внедрила приоритизацию трафика Google Cloud для безопасной миграции критических нагрузок

Uber uses Google Cloud to prioritise critical traffic Thu, 27th Aug 2026 Uber has deployed Google Cloud's application awareness on Cloud Interconnect across parts of its network to support workload migration to Google Cloud. It first used the networking feature in private preview before deploying it on Google Cloud Interconnect links in Phoenix, Arizona, and Ashburn, Virginia. The system lets Uber classify traffic and prioritise user-facing and other critical applications over less time-sensitive data transfers. The set-up addresses a central problem for large hybrid and multicloud operators: how to move large volumes of data without disrupting services that customers and internal teams rely on. For Uber, that challenge has become more urgent as analytics workloads and AI-related data flows place heavier demands on network links. In conventional interconnect set-ups, companies often treat all traffic the same or add spare bandwidth to handle peak demand. Uber said that approach was too costly and too uncertain for a global platform managing major traffic swings and large-scale data movement. Traffic priority Google Cloud said application awareness on Cloud Interconnect allows traffic to be sorted into six classes instead of being handled on a first-in, first-out basis. It uses DSCP marking and queuing profiles so operators can assign strict priority or bandwidth-sharing rules to different types of traffic. That allows business-critical services to be protected during bursts of demand or periods of congestion, while lower-priority transfers wait. The goal is to keep latency more predictable for time-sensitive applications and reduce the need to buy excess network capacity purely as a buffer. Uber said the immediate business effect was greater control over which applications take precedence during major events, planned or unplanned. It also said the change helped unblock the migration of significant workloads to Google Cloud by lowering the risk of service interruptions during switchovers. The company presented this as part of a broader effort to reduce operational overhead while maintaining service reliability. Hybrid AI, analytics and distributed applications can require heavy data transfers across on-premises and cloud environments, increasing the risk of congestion if traffic is not managed closely. Migration impact For Uber, the networking set-up became part of the foundation for moving more strategic workloads to Google Cloud. The company said the ability to shield important traffic from congestion gave it more confidence to proceed with migration work that might otherwise have carried greater operational risk. Google Cloud also described Uber as an early design partner for the product, suggesting the feature was shaped in part by the demands of a large global operator. That indicates the deployment was not simply a standard customer installation but also a test case for how cloud providers are adapting interconnect services to more complex enterprise traffic patterns. Cost control is another theme in the deployment. Rather than overprovisioning bandwidth for the most extreme peaks, Uber said it can use existing Cloud Interconnect capacity more efficiently by aligning traffic handling with expected network needs. That matters because AI and data-intensive computing can create brief but substantial traffic spikes that do not always justify permanent increases in network capacity. Prioritisation gives operators another way to protect services without relying solely on larger links. Across the industry, cloud providers and large enterprises face a similar tension between rising demand for AI infrastructure and the need to preserve dependable application performance. Moving data between systems, regions and environments is becoming more common, but network congestion can quickly undermine migration plans or affect customer-facing products. Uber said its experience offers a model for other organisations dealing with hybrid cloud complexity. It identified three main results from the deployment: preserving business continuity, making more efficient use of bandwidth and clearing the way for larger workload migrations. Harry Liu, Director of Engineering at Uber, described the deployment as central to that process. "Application awareness on Cloud Interconnect was the key that unlocked our ability to migrate more strategic workloads to Google Cloud and is critical for maintaining service reliability during peak global demand. By allowing us to intelligently prioritize traffic, it helps us ensure that we can protect our higher priority services and make our infrastructure more efficient, lowering our total cost of ownership. This wasn't just a feature deployment; it was a deep engineering partnership that delivered a solution critical to our business," Liu said.