 Yes. Hello everyone. My name is Kong Ha Min from Vietel Group. Today I and my colleague will bring for you an infrastructure management topic. Yes. Today we will talk about why we need an infrastructure management system and how we use open source solution network to build our own infrastructure management system. Yes. For the first, I will introduce a little about my company, Vietel. We are a big telco and IT cloud and a telco provider in Vietnam other 10 countries in the world. So we have many customers to provide telco and IT service to many customers in the world, about 100 billion customers. We need a very big infrastructure to deploy our IT and telco service that we have many things to management such as from physical component as server suite router to a network component such as network address, mark address or some network connection. And we have also have many logical cluster component such as cloud storage, load balancer pairing. So in this reality, we have to manage many things in our infrastructure. So we have many elements on many sites. First, we need to use to have an information management infrastructure information management system to audit, to check how many network infrastructure we are using and how many left. For example, we have many right but some right are used and we need to check how many space available for us and how many left. For example, another example is IP management. We have many prefix but some prefix has you but some not you. Then we want to check how many IP available in each prefix so we can use this available IP to create new VM to deploy new my new IT service. So the second demand that we need to upgrading our infrastructure in single place. So then other systems in our ecosystem can use this infrastructure management system to add input data for monitoring, for auditing, for alerting. So for example, that's when a VM in all infrastructure has problem that the alert system, the monitoring system need to assess this single infrastructure management system to collect what the owner, the manager of this VM is and the location of this VM. So can it alert the VM manager so he can see that the VM is down and he is responsible to assess to this VM and fix and bring up it again. So the other demand is to check the different between our design state with the current state. For example, we can have some VM has duplicate IP in the environment due to human error, this conflict so that we need infrastructure management system to detect it and then alert to VM manager, object manager to fix it to correct that no VM in all infrastructure has duplicate IP as well as the system work correctly. So for the three requirements as I talked to you, we had to evaluate many solutions. The first solution is we can buy a commercial solution from a provider such as the single solution is we buy, we can build our own solution from scratch but it's very cost many human and money to make it. Then the first solution for us is based on open source solution and then add all need to extend the solution to meet our requirement. So after evaluating the solution we decide to base our infrastructure in management system on networks. It's a very popular solution infrastructure management system currently. So why we chose it? First, network currently provide for our company almost the feature we need to manage our infrastructure system. So the second reason is the network has an open source solution with an open design. So our company easily extend it to add our need feature. It's based on the Django and it has an edit TTP REST API so we can integrate our system or other system such as monitoring system, audit system and alert system to the network to provide input for other system. So it has many features our company need such as it can manage device, manage IPRS, manage the right reservation so we can use it for monitoring for auditing, for capital liberty planning, with a visit for device for physical component for IPRS but with a big company like us we need a more HR requirement feature for our daily activity monitoring on operating our infrastructure. So what is the first? Network only provides a solution for many basic device components, a single device or a right chassis device but currently it cannot manage the logical component so that's cloud storage system or a pairing load balancer so our company needs to manage this. The second, sometimes we need some more data from real world. We need to collect to provide input for the maintenance and operating activities. Network do not have an automatic recovery service for them. For the third requirement is sometimes our company need to validate if the real state and the design state in our infrastructure are different so I talked earlier we can identify and let the owner manager of this object can fix this to bring it in our design architect. So to show these three problems that's our solution is the first we extend network to add more capability to manage the logical cluster object so that we add more the Django class and API to manage these objects in the infrastructure. So the second we add some feature to manage the HR object to manage the data from automation discovery system to put it in the web service so we can collect it and when network operator need it it can access to the web service to see it and you see it to fix the problem when happened. So the third we build a discovery service. The service we will collect data from real object from object infrastructure and collect it and put it in the web service. And the second feature of the discovery service is compare the design state the network manager put in the web service and compare with the real state in the environment and when it detects different it's then allowed to the object manager to show the object manager and the network operator can detect it and the fix is to bring back the real state to the design state. So what is the feature we add to network to complete the feature we need in our company that's the first thing you can see is that we had some new property to manage the load balancer pairing because it can have two or three load balancer devices in a pairing and we add a new object called load balancer to manage them and we can manage some HR information from this such as Vip and Poi or we can manage more in its quick route that many operators of our company need them when troubleshooting an issue infrastructure problem. So the second feature we add is detects the mistake between the real state and the design state. As you can see that we have two VMs with duplicate IP when the discovery service detects us it can trigger an alert to manager and the operator in an email. So when they have emails then they can access to the device and fix it. So it is the feature we add to network to manage our infrastructure management system. So this is my presentation. Thank you. So do they have any questions or anyone use network to manage the infrastructure? Yes, yes, yes. My current implementation on our company is I add using our fork from network call and when an upstream has update we will manage. So it gives me more flexible to customize network to adapt our requirement but because network plug-in sometimes it cannot have enough flexible for us. But we will have more time to manage code from upstream. So we must to design our code to not conflict with network call. Yes, another question. Do you have any questions for me? Thank you very much.