Infrastructure

Introduction

This document illustrates the details of our Chat-Tonic platform, which is consumed through a PaaS scheme (platform as a service), so the details of the implementation of some parts of the solution are proprietary to Chat-Tonic, and are beyond the scope of this document.

Infrastructure

Most of the platform is hosted on its own virtual servers, whose provider is Linode. The APIs as well as the Chat-Tonic, Customer Service, and Web Chat-Tonic frontends are fully managed with the described scheme.

Servers are periodically updated by our sysops team, especially against recognized public vulnerabilities.

Our solution has multiple backup instances to guarantee that the data is protected. Our server provider performs full server backups on a daily basis. It would be possible to restore a copy directly from there in case of total server failure.

In addition to this plan, the data is stored in a variety of services, but mainly everything related to the management of the chatbot is stored in its own internal database. This data is backed up daily, and checks are made on these backups to be able to restore the data periodically at least once every 6 months. The rest of the infrastructure, in case of a total hardware collapse, it is possible to restore directly from our code repositories. We keep the last 30 copies of the database for any contingency.

We have an internal test environment, where we carry out tests on new features of the platform. This environment is not normally available to customers. If it is necessary to have a test environment to test new functionalities of the bots, it can be done directly in production, in an environment separate from the production one.

We currently have deploy our production platform in the USA (Central) and the EU (Frankfurt), which the client can choose when launching their chatbot if they need to comply with different security requirements or legal regulations. This cannot be changed once the chatbot is launched.

Architecture

This is an architecture diagram, considering Chat-Tonic as a PaaS whose connection to the outside world is solely through APIs. This representation abstracts within Chat-Tonic the concepts of handling conversations, messages, leads, stats and referral to human, among others.

We have created our own chat-bot platform from the ground up. We have built and have control of the platform from end-to-end. Our main core is an API that is able to handle all the incoming traffic of the different chat platforms, to then be interpreted by the Chat Bot Engine and come up with the answers to the final users. We also provide a separate API to handle all our Web Chat-Tonic platform, as well as our Customer Service (live agent) platform.

All these are implemented using the latest frameworks from:

  • Node.js

  • MongoDB

  • RedisDB

  • React

Our structure is to host all clients in a multi-tenant environment. The environment could be based on USA or EU depending on the client’s requirements. In special cases, we are able to host a client in a separate private cloud, managed entirely by us, that should be discussed during the commercial agreement

Security (see complete report in a separate page)

 

Help and support

Our support is through Trello, where both the developer assigned to the project and the project coordinator are alerted to any changes that may occur there. In this way we are alerted if there is a ticket to attend. Additionally, you can use the internal ticket reporting tool from our Dashboard where those generated tickets feed Trello itself, as well as the progress status of them.

Our support hours for bots inquiries or reports is from Monday to Friday from 9 a.m. to 6 p.m. (GMT-3)

Any interruption of the service that you may experience, can be reported by the means that I mention before, outside the mentioned attention hours, and we normally respond within 6 hours of generating the ticket.