As recent as last month, the Red Hat name has just solidified its foothold in Asia by launching their very fist Innovation Labs in Singapore. The third chapter of the open source champion proven Innovation Labs initiative is built and launched in the bid of serving the rapidly growing Asian market. The choice of launching the Innovation Labs in Singapore is down to the fact that Singapore is technically the economic hub of the South East Asian region. Singapore also serves as their South East Asian headquarters making it the ideal place to have a full-fledged Innovation Lab for the Asia region. Red Hat’s Innovation Labs, according to Damien Wong, Vice President and General Manager of Red Hat ASEAN is not just limited to one location though. The Linux for enterprises provider can, in theory easily set up its Innovation Labs in any given space. Known as a pop up lab, Red Hat’s clients do not have to go all the way to Singapore to get a glimpse of how Red Hat works.
We recently had an exclusive chance to speak to Damien Wong and Massimo Ferrari, Red Hat’s Management Strategy Director at their Forum 2017 in Malaysia on the major milestone that is the Innovation Lab in Singapore and also the latest addition in their enterprise solutions family; Ansible Automation. In this second part to a three-part series of the interview we spoke to them about their latest addition to their wide repertoire of solutions, the Ansible Automation and the concept of the Ansible Tower.
Firstly, the Ansible Automation platform is made up of two moving parts; the Ansible Engine, and the Ansible Tower. These two parts make a recipe of automated tasks that helps an organisation save resources like time, money, and man hours. The core of Ansible Automation is its Engine which contains what is known to Red Hat as the Playbook. It is a series of “If This Then That” sort of codes that should automatically run the most menial of tasks. In a sense the Engine is supposed to streamline individual processes and make them completely automatic and mindless. Then there is the Ansible Tower, another essential part of the Ansible Automation program. The Ansible Tower is likened to a central control. It the Playbook’s control panel where each Engine can be turned on and off and customised from the Tower itself. It is described as a controller of some sorts, an administrator to all the Engines. It’s purpose? Solely to make monitoring and regulating the Ansible Engines much easier; a sort of one stop shop for Ansible Automation program.
Next we spoke about the integration of the Ansible Automation program with other moving parts in Red Hat’s arsenal of enterprise solution programs. For example, the integrating Red Hat insights with Ansible Automation. In essence, the insights is Red Hat’s version of what machine learning could do. The algorithms implemented learns, analyses, and ultimately suggests possible or best solutions and improvements for its users. All in all it is a pretty similar system in description compared to other machine learning systems out there. According to Red Hat, insights is mostly an analytics software that does more analytics and suggestions for future preventions and improvement than anything else. An integration of the Ansible program with insights alone could be limited in terms of only automating the most common of processes. The point of Ansible is technically to automate the most mundane and routine of processes to create a higher level of efficiency and effectiveness in an area.
The Ansible Tower and Engine though is designed to work and integrate with all Red Hat’s services. The big idea here is to provide a complete solution to the client. The Ansible Automation program is designed to work with more than just insights itself. Ansible’s Tower itself is designed to work with even more complex combinations of the Engine itself. For example (going back to insights) Ansible can be paired with insights to not just recommend solutions and best actions. The algorithm can be taught to undertake preventive actions on its own through a series of chained commands from within the scope of the Tower itself. In that sense an organisation can complete automate certain processes that would free up their own staff and other resources to innovate even further. The ultimate goal is to make life better for us, the end consumers by giving us better products and services.
Then we come to Red Hat Insights as a specific Machine Learning solution for enterprises. Just like plenty of the machine learning algorithms out there the insights learns patterns and its users’ behaviour to improve and add to its own algorithm. The difference here is that insights benefits from the wealth of data that Red Hat, as an open source solutions provider has collected over the years regarding their clients. Of course there are certain data that Red Hat will not use or disclose on its collective insights databases for its customers. Another difference with insights, according to Red Hat is that the algorithm changes all the time. All changes are moving towards the refinement of the technology though so everything is beneficial to the client. But that also means the insights is not just a single algorithm, rather it is a collection of multiple algorithms that it has learnt over the years of its development. As long as the insights project goes, it will add to its collection of algorithms and find new ways to analyse data depending on its customers.
How insights works is quite unique compared to other machine learning algorithm out there. Of course the fact that this particular Artificial Intelligence (AI) is designed purely for enterprises contributes to its unique form and function. It works with a collective set of data that has been collected both in-house and through participating clients. It compares the data of its current clients to the collective data sets to recommend what might seem to be the best or correct solution. The data collected from this particular client then is added to the collective database. Despite all the provided data, Red Hat still admits that some of its solutions may not be the best at a given time. It is trainable though to suggest the best possible solution when required. The data sets that Red Hat collects into its collective data base are mostly infrastructure type data though; specific enterprise information are usually sensitive.
There are three types of use cases and configuration for the Red Hat insights – security, compliance or configuration purpose, and performance. In the case of security implementation insights can run across its client’s database and cross check every file with the most current public safety databases for malicious or harmful files and software. It will then recommend steps that are to be taken to keep your systems clean and safe. In the case of configuration or compliance use case insights works as a sort of brain. If figures out what all systems that are fed and connected to it does. It learns the workings of the programs that are connected to it, what makes them work, and if there are any inefficiencies the system programming. From there it will form or recommend a process flow or it will match and create an effective workflow for its clients. In the performance domain insights understands and finds your physical infrastructure alongside the software implementation. From there Insights basically does all the work in while understanding how the systems work and calculates its risks in the future. It will, of course then ask users what they want to do.
The Red Hat Insights is built upon the aim of serving the enterprise clients better so that they can improve their service to their customers, which could be us. It is not a one-fits-all solution as well unlike the Google Deepmind framework and IBM’s Watson. Red Hat though can customise the insights for its various clients as they do understand that each client is unique and requires different approaches and have different requirements. It is a more focused program that is designed to help its enterprise customers more than anything else.
Source: Red Hat Malaysia
Also published on Medium.