Re: Defining load balanced, failover clusters for DID Document serviceEndpoints?

...where executable code is versioned, packaged, and distributed as signed Verifiable Credentials (aka signed Structured Credentials), executed on a random collection of load-balanced Compute Nodes scattered across the Internet (defined by Structured Credentials) with both input and output parameters also defined by signed Structured Credentials.

Michael Herman

Get Outlook for Android<https://aka.ms/AAb9ysg>
________________________________
From: Michael Herman (Trusted Digital Web) <mwherman@parallelspace.net>
Sent: Friday, January 7, 2022 9:43:37 AM
To: David I. Lehn <dil@lehn.org>
Cc: public-credentials (public-credentials@w3.org) <public-credentials@w3.org>
Subject: Re: Defining load balanced, failover clusters for DID Document serviceEndpoints?

Dave, are you aware of any "normal scalable infrastructure best practices" that aren't based on some sort of centralized routing infrastructure (e.g. round-robin DNS as a simple example)?

Ultimately I'd like to find/build a pattern for creating "a fabric of fully decentralized compute nodes" that still works within/compliant with the current constraints of the DID-CORE spec.

Michael

Get Outlook for Android<https://aka.ms/AAb9ysg>
________________________________
From: David I. Lehn <dil@lehn.org>
Sent: Friday, January 7, 2022 8:22:00 AM
To: Michael Herman (Trusted Digital Web) <mwherman@parallelspace.net>
Cc: public-credentials (public-credentials@w3.org) <public-credentials@w3.org>
Subject: Re: Defining load balanced, failover clusters for DID Document serviceEndpoints?

On Fri, Jan 7, 2022 at 5:24 AM Michael Herman (Trusted Digital Web) <mwherman@parallelspace.net<mailto:mwherman@parallelspace.net>> wrote:
Has anyone else thought about this? ...for example, the value of serviceEndpoint can be a map or set of URIs (https://www.w3.org/TR/did-core/#services). This collection of URIs can serve as the list of serviceEndpoint/agent instances in a round-robin or load-balanced cluster. Second, the service type could be "clustered" to signal this is explicitly a load balanced, fallover-enabled serviceEndpoint/agent.

Thoughts?


Implementation details for a web endpoint are out of scope for the specs defined here.  Use normal scalable infrastructure best practices for your services.

-dave