SVTA_LOGO_FINAL_revised_2022

Edge Meeting Minutes
March 12th, 2026 9:00am to 10:00am PT</p

Attendees

Yoav Gressel (Streaming Video Technology Alliance (SME))
Sanjay Mishra (Verizon)
Alan Arolovitch (2you.io)
Brian Haberman (Fastly)
Chris Lemmons (Comcast/NBCUniversal)
Alfonso Siloniz (Telefónica)

Minutes

  • Summary

The meeting focused on discussing orchestration approaches for hosted caching, with Alan presenting his takeaways from recent sessions. The group explored the challenges and requirements around Kubernetes implementation, including the need to support both VMs and container-based workloads while maintaining interoperability between different hosting providers. They discussed the separation of control plane and provisioning components, with agreement that common APIs would be needed for user management and inventory across different platforms. The team also touched on the potential need for reservation and scheduling capabilities, though this was identified as a future consideration rather than an immediate requirement. The discussion concluded with an agreement to create two documentation tracks: one for control plane implementation blueprints focusing on OpenStack and Kubernetes, and another for provisioning APIs that would work across different hosting environments.

 

  • Action Items
  1. Alan: Start a document outlining control plane implementation blueprints, focusing initially on Kubernetes and OpenStack, describing possible architectures and approaches.
  2. Alan: Start a second document specifying the required provisioning APIs (including user management, inventory, and RBAC), with the goal of enabling interoperable tenant access across multiple hosting providers.
  3. All participants: Review the two documents (control plane blueprints and provisioning APIs) in detail and map out requirements in upcoming calls.
  4. All participants: Dig deeper into resource allocation and inventory, especially as related to locations and deployment options, for future meetings.

 

  • Antitrust Statement & Orchestration Strategy and Standardization
  • Alan and Yoav discussed the challenges and considerations around orchestration, particularly in the context of Kubernetes and other options like VMs.
  • They explored the need for a frictionless approach that could be adopted beyond early adopters, considering the differences in implementation across public clouds.
  • The group agreed that while Kubernetes is a key option, they must also consider other methods, such as OpenStack for VMs, to meet the diverse needs of tenants.
  • They identified two main areas for standardization: the workload control plane and provisioning, which includes user management and inventory.
  • Alan suggested crafting a scope of work for these areas, focusing on creating common APIs that would work across different cloud providers.

Cross-Platform Orchestration Architecture Planning

  • The team discussed how to handle orchestration across different platforms like Kubernetes, OpenStack, and VMware.
  • Yoav and Alan agreed that two main documents would be needed: one describing possible Kubernetes architecture implementations and another specifying shared APIs that hosting providers would expose.
  • They also discussed the need for a framework similar to AWS CLI for provisioning and workload management, with RBAC and IAM capabilities to be defined.

User Access and Inventory Standardization

  • Alan and Yoav discussed standardizing role definitions for user access control, agreeing on a predefined set of roles to be transferred through standard authentication and authorization channels.
  • They also explored the complexity of inventory management, particularly regarding the discovery of locations and the need for multiple APIs to handle different deployment scenarios.
  • Alan highlighted the lack of a reservation mechanism for scheduled workloads, suggesting this could be an area for future development to support tenant scheduling needs.

Kubernetes Resource Allocation Strategies

  • Yoav and Alan discussed cluster management and resource allocation in Kubernetes and Google environments.
  • They explored the possibility of abstracting resource reservations away from specific containers or pods, allowing for more flexible allocation of CPU and memory resources.
  • Alan suggested that this could be achieved through a control-plane-agnostic approach, potentially supporting features like namespace limits.
  • They agreed that building a glue layer between Kubernetes and OpenStack might not be practical, and instead focused on recommending approaches rather than prescribing specific configurations.

Cloud API Interoperability Requirements Discussion

  • The group discussed API support and interoperability between different cloud providers, with Chris suggesting they start by defining minimum requirements before considering full FCI capabilities.
  • Alan expressed concerns about over-engineering and proposed creating two documents: one for control plane implementation focusing on OpenStack and Kubernetes, and another for provisioning APIs.
  • The team agreed to avoid prescribing unnecessary features and focus on enabling tenants to work with multiple hosting service providers using the same toolset.
  • Yoav raised questions about resource allocation in Kubernetes, and the group briefly discussed the potential use of functions, though they decided to postpone defining specific function requirements until a clear need arises.

 

Next Meeting: Thursday, March 26, 2026 9am PDT

Have A Question ABout Membership?

Schedule A Meeting

[hfcm id="18"]

Send An Email

Don’t want to schedule a face-to-face meeting just now? No problem. Simply send your membership question to info@streamingvideoalliance.org or fill out the form below and someone will get back to you as soon as possible.