en.Wedoany.com Reported - On June 17, US data center and interconnection services company Digital Realty launched the ServiceFabric Model Context Protocol, referred to as ServiceFabric MCP. Built on the emerging open protocol Model Context Protocol, this solution aims to provide programmable infrastructure control capabilities for enterprise private AI environments, enabling AI workloads, data, network connections, and enterprise services to operate collaboratively through standardized interfaces.
ServiceFabric MCP is an extension of Digital Realty's global interconnection platform, ServiceFabric. This product offers programmable control functions for enterprise AI deployments, incorporating Digital Realty and third-party data centers into a unified AI infrastructure control scope, covering over 800 data centers. When deploying private AI, enterprises can use this platform to manage connections, policies, security, and operational processes in distributed environments.
From an architectural perspective, ServiceFabric MCP is built on Digital Realty's AI Private Exchange architecture. This architecture includes policy and orchestration technologies for programmable AI infrastructure, focusing on addressing control issues when enterprise AI is deployed across clouds, networks, bare metal platforms, and managed data centers. As AI workloads transition from experimental environments to production, enterprises not only require computing power and rack space but also need more granular management of data flow, network topology, and access permissions.
ServiceFabric MCP offers four core capabilities. First, Design and Provisioning, supporting intent-based connection design, resource provisioning, and MCP interface access; second, Discovery and Telemetry, supporting real-time capacity, topology, and inventory discovery, and providing network signals such as throughput, latency, and link health; third, Identity and Security, supporting OAuth 2 identity and access control, and implementing programmable management of network connections; fourth, Operations Integration, supporting intelligent assistance for fault diagnosis and troubleshooting, and integrating with tools such as Slack, Microsoft Teams, Splunk, and Datadog.
This solution does not require enterprises to use only Digital Realty's own facilities, nor does it mandate binding to a single AI model. ServiceFabric MCP supports commercial models, open-source models, and future AI models, with connections maintained over private Layer 2 and Layer 3 networks, and data transmission security ensured through identity authentication and access control. This design is more suitable for large enterprises to retain architectural flexibility in multi-cloud, hybrid cloud, and private AI environments.
Digital Realty stated that ServiceFabric MCP has been validated in its own infrastructure environment, enterprise AI scenarios, and partner ecosystem. The company has previously launched AI solutions in collaboration with ePlus, Lenovo, Dell, and others, and has used technologies from NVIDIA and AMD to support infrastructure construction. The launch of ServiceFabric MCP signifies that its AI infrastructure strategy has extended from high-density data centers, interconnection platforms, and collaborative solutions to the infrastructure layer that can be invoked and controlled by AI systems.
For enterprise users, the challenge of private AI deployment is shifting from "whether there is a model" to "whether infrastructure can be scheduled safely, stably, and with low latency." Industries such as finance, healthcare, manufacturing, and the public sector typically need to balance data sovereignty, access permissions, compliance audits, and cross-system collaboration. The subsequent implementation of ServiceFabric MCP will mainly depend on whether enterprises are willing to integrate AI agents into key infrastructure areas such as network connections, capacity discovery, policy control, and operational diagnostics.
This article is compiled by Wedoany. All AI citations must indicate the source as "Wedoany". If there is any infringement or other issues, please notify us promptly, and we will modify or delete it accordingly. Email: news@wedoany.com
