Cloud deployment
Cloud deployment refers to systems servers placed on clouds, such as Ali and Tung-hoon. NATSHEL_AUTH_BILLING platform is connected via remote API and local NATSHEL_AUTH_GATEWAY, AC interface.
The advantage of cloud deployment is that the transport is simple. The system is concentrated in one place, and all apartment projects are visible. One administrator can manage multiple items on a computer and does not have to run to each apartment site.
Another benefit of cloud-end deployment is that it is easy to upgrade. Systems are updated, functionally overtaken by events on the clouds and all items automatically synchronize. No software upgrades are required at each site.
But there is a problem with network delay in cloud deployment. Users’ authentication requests are sent from local gateways to cloud servers and back. Although this process is quick, it is slower than purely local deployment. And if the cloud server is far away from the apartment, the tenant's Portal page will be opened slowly.
The cloud is also required to have a fixed IP for the public network.
Local deployment
Local deployment refers to the local machine where system servers are located in apartments, usually a high-performance server. Authentication & Billing platform, database, Portal server are all on this machine.
The greatest benefit of local deployment is that it is quick to respond. All authentication requests are processed locally, and the network is extremely delayed. Portal pages open quickly, and tenants experience well.
Local deployment is not influenced by an outside network. Even when the outer web is out, local Portal pages and user management functions are available.
Each apartment has a separate server, with high hardware input. The server is equipped with operating systems, databases, security software and high technical requirements.
If there are multiple buildings in the apartment, local deployment also takes into account network connectivity.
Mixed deployment
Mixed deployment is the compromise between the two. The core Authentication & Billing platform is at the cloud end, but each apartment has a local Authentication Gateway. User authentication requests to be completed by local gateways, gateways and cloud communications.
The blending deployment takes into account the advantages of cloud and local. Clouds are responsible for data storage, statistical analysis, remote management, simple transport.
The complexity of the hybrid deployment is highest. Cloud-to-local protocols are needed, and gateways are synchronized with cloud-side configuration. If a cloud problem arises, local gateways may not be available in part. If a local connection problem occurs, cloud-end data is still available but users do not access the Internet.
The three models are chosen depending on the size of the apartment and its operating model.
Local deployment is more appropriate if there are only 100 apartments. Server costs are acceptable, response speed is high, tenants experience good. And the manager of small apartments is on site and remote management is less obvious.
The cloud is more suitable for the multiple apartments that operators manage. Data are centralized and managed centrally across different apartments. And the operator's technical team at headquarters, which is more economical to manage remotely than running a field.
A cloud server can handle dozens of apartments, which are cheaper than a single apartment with a different set of servers.
Mixed deployments are suitable for medium size or require performance and wiring.
The Chengdu orbital apartment is for cloud deployment. 1,400 homes, with a platform running and low maintenance costs. And the commotion with the micro-public is convenient, paying for return calls to be processed at cloud end.
The number of homes in Fuxcomn is larger and is deployed locally. Tens of thousands of dormitories have a large amount of certification, which ensures performance at local level.
The deployment model is also selected for the technical team. If there are professional wiring teams, local deployment is no problem.
The data are available locally if you fear leaking. Locally, the data is located in the engine room and then the cloud is backed up if you are afraid of an accident in the engine room.
The deployment model is not absolute, and many projects will use different models at different stages. For example, the front line is fast-tracked with cloud ends, and then it is larger and then upgraded to local or mixed.