Oracle Cloud · AWS, Build a Multi-Cloud “High-Speed Private Network”… AI Workload Bottlenecks May Decrease

Oracle (ORCL) and Amazon Web Services (AWS) have set aside some differences in their long-term competitive landscape and are collaborating to address the “multi-cloud” demand. As more enterprises use different cloud services, both parties have decided to build a high-speed dedicated network to essentially connect the two platforms into a unified computing environment.

The core of this announcement is to provide a private, high-speed connection between Oracle Cloud Infrastructure (OCI) and AWS cloud. Their joint customers will be able to migrate data or run applications without passing through the public internet. This move aims to reduce latency while enhancing the stability and security required for enterprise workloads.

This reflects the recent reality that “multi-cloud” is no longer a choice but a fact in enterprise IT environments. Companies often do not rely on a single cloud but deploy databases, applications, and AI training and inference systems across different cloud platforms. The challenge lies in connecting these clouds, which is both complex and costly.

Previously, enterprises had to assemble external network operators, manually configure connections, and invest in expensive physical equipment to build their own connectivity architecture. This process often encountered the so-called “data gravity” problem—data being effectively bound to a specific cloud, making migration difficult and costly.

Such issues have become more prominent with the proliferation of AI. For example, companies can adopt a “disaggregated stack” architecture, running high-performance databases on Oracle OCI while training models and handling application logic via AWS SageMaker. But without high-performance connection channels, latency between environments would increase, and actual service utilization would inevitably decrease.

To address this bottleneck, Oracle and AWS will enable interoperability between Oracle’s proprietary interconnect services and AWS’s multi-cloud interconnect standards. Customers will no longer need to design complex manual routing or data replication schemes. Operations such as running applications on one cloud while storing data on another, or migrating all data from one platform to another, are expected to become simpler.

The market views this move as a symbol of “co-opetition.” Senior analyst Rob Strech of research firm TheCube Research pointed out that multi-cloud is no longer a strategic experiment but a practical operational reality for most enterprises. He noted that reducing network complexity between the two major cloud platforms will greatly facilitate the construction of enterprise AI architectures and disaster recovery systems.

Especially in the AI era, data storage locations and model execution sites are likely to be separated. In this architecture, network bottlenecks could impair AI performance and increase costs. This collaboration not only introduces new connectivity services but also offers an opportunity to flexibly reshape enterprise AI infrastructure design.

This is not an entirely new trend. Google Cloud launched cross-cloud interconnect services in May 2023, later expanding dedicated private connections with AWS. Oracle has also established interconnects with Google Cloud and Microsoft ($MSFT) Azure. However, industry commentary suggests that its direct integration with AWS is more symbolic.

Oracle Product Management Senior Vice President Nathan Thomas explained that this release is an extension of the existing “Oracle Database@AWS” collaboration. He stated that joint customers will be able to leverage this to modernize applications, integrate data, and expand opportunities in generative AI.

This high-performance connection targets large-scale enterprise workloads and is planned to launch within the year. The initial service region will be AWS US East (Northern Virginia). While cloud market competition will continue, at least in the multi-cloud and AI infrastructure fields, this case demonstrates that “interconnectivity” is becoming a more critical competitive advantage than “closed ecosystems.”

TP AI Notice: This article is summarized based on the TokenPost.ai language model. There may be omissions of main content or discrepancies with actual facts.