AirDeck Architecture

AirDeck is a software stack that enables enterprises to deploy and operate scalable private clouds. It is developed by combining open-source and proprietary software, and designed to run on off-the-shelf white-box servers without any specialized hardware. The stack is composed of four core layers:

 

Doppler

Distributed and scalable cluster management

It is most cost effective to build a cloud infrastructure using off-the-shelf servers, to be a viable enterprise solution however, it requires robust software to maintain high availability on a wide range of commodity hardware at all time. Doppler is AirDeck's solution to deliver enterprise class high availability. It is based on Kubernetes, which provides basic scalable cluster management, and enhanced with sophisticated error handling and failure recovery capabilities. Unlike other solutions that only try to address single point of failures.

Doppler implements a comprehensive distributed monitoring system that covers every aspects of an environment including hardware components, network connectivities, and software services, as well as a set of error detection and isolation algorithms to maintain high availability and minimize disruptions to user applications at the same time.​

overview-appllo-03.png
 
333.png

Apollo

Powerful scale-out software defined storage

There are a number of scale-out software defined storage solutions available, the common problems users encounter with these solutions are poor performance and limited scalability. When performance and scalability are in demand, users have no choice but to resort to specialized storage systems (both hardware and software). This undercuts the cost effectiveness of utilizing commodity hardware, especially considering the raw capability of the hardware (such as NVMe), is largely untapped. 

Apollos is AirDeck's solution to these problems. It is designed to fully leverage storage capability available in off-the-shelf servers to deliver a powerful scale-out software defined storage. Unlike any other solution, Apollo is the first storage system developed with the versatility to provide software-defined storage on demand in three different dimensions: capacity, redundancy, and performance. With this, you can always provision the right storage for your applications. There is simply nothing more powerful existed today!

444.png
 

Houston

Intelligent resource scheduling and

dynamic load balancing

Most of the private cloud solutions rely on operators to manage and allocate resource in provisioning virtual machines, auto resource scheduling and load balancing are advanced add-on features that also require complicated setup if available. Opposite to the manual labor, Houston is AidrDeck's solution to off-load users from tedious manual operations and provide an auto-pilot operational environment.

It collect information of overall resource utilization on a regular basis and formulate intelligent insights to an operational environment, and whenever necessary it will always distribute or adjust workload in an environment to achieve the optimal balance of workload performance and effective resource utilization. 

 

 

houston-sifi-nobg-01.png
 

Arc

GUI and API gateway to the cloud

There are a number of scale-out software defined storage solutions available, the common problems users encounter with these solutions are poor performance and limited scalability. When performance and scalability are in demand, users have no choice but to resort to specialized storage systems (both hardware and software). This undercuts the cost effectiveness of utilizing commodity hardware, especially considering the raw capability of the hardware (such as NVMe), is largely untapped. 

111.png

Arc

GUI and API gateway to the cloud

There are a number of scale-out software defined storage solutions available, the common problems users encounter with these solutions are poor performance and limited scalability. When performance and scalability are in demand, users have no choice but to resort to specialized storage systems (both hardware and software). This undercuts the cost effectiveness of utilizing commodity hardware, especially considering the raw capability of the hardware (such as NVMe), is largely untapped. 

222.png