Task scale models

Scaling at the datacenter level and transition to SuperKANT.

Greetings, colleagues! We continue to share with you the articles, which in essence are a practical example for building a mathematical model and working out adaptive consensus in GONT. Today we are talking about the task scaling models.

GONT consists of both a variety of gVM micro-cores (GONT.KANT level) and a multitude of micro-servers (individual boards) – the SuperKANT level.

In general, all gVM cores can interact with each other at any GONT level! Therefore, we come to the following interaction model:

All-to-all server-to-server communications

The smart contract model in GONT implies that any two gVM machines in a distributed GVM container interact in the same way (no matter how far along the routing levels in the container they are located).

At the same time virtual transaction channels between cores should work without delays (congestion for transactions).

Solution scalability

The Goya chip must be designed to support unlimited GONT scaling at the datacenter level.

Scaling also means that the interaction between Cluster1 and Cluster2 (inside one GOYA chip) will be at the same bandwidth as the interaction between Cluster2 and Cluster3 (between GOYA chips).

Global scaling of computing tasks at the datacenter level within the framework of the SuperKANT concept.

GONT boards:

 

Thank you for attention! See you in the next articles.
GONT

Leave a Reply