Fog Computing
The future of "IoT" landscape is moving towards "Fog Computing". Fog computing is a decentralized infrastructure where the bridging with devices, data analytics, computation and "first response" applications are deployed in a more logically, physically close or isolated from the cloud. Fog computing is basically coming out the cloud but still persisting the implementation and functionalities of cloud and "Internet Of Things" in a more effective and efficient way. Fog Computing is the more generalized name or the industry level name of Azure IoT Edge.
"Fog Computing" High Level Architecture
Azure Stack
Azure stack is the replica of the Azure cloud environment in local or on-premises. To be more precise; it is the Azure hosted in your private data center or in a more remote location. Azure stack is a Integrated system of both software and hardware components deployed in your local. Azure stack enables IT's too develop, deploy and maintain some of the Azure resources within the on-premises.
Why Azure Stack in Azure IoT Edge
Based on the above "Fog Computing High Level Architecture" Diagram. There is a intermediate physical or logical layer which lies between the device and the public cloud. This layer should be "Azure Stack" to my assumption . Azure fulfills all the rules and constraints of the "Fog layer", it will play a major role in the decentralization or intermediate layer of the Fog computing.
Letz check why it will play a big role in the Azure IoT Edge platform.
Letz check why it will play a big role in the Azure IoT Edge platform.
1. Reduction of high network dependency, continuous execution of the systems even in "Out of network" scenario.
To the question from the previous post "Where the container will be hosted ? will the container be hosted in the cloud or in the on-premises" Azure Stack would be right choice to "answer". The Container's will be hosted in the Azure but on-premises. This design or implementation would reduce the unwanted network usage and effectively reduce the time for presumptive analytics to be done. One of the major advantages of Azure Stack is it capabilities to perform in disconnected mode.
2. Miniature cloud platform even in remote plant location.
Azure Stack is the replica of Azure environment. This will allow to design the same architecture for both the cloud and the on premises application. Once the Azure Stack is created in the remote plant the things implemented in Cloud can be easily moved or replicated in the Azure Stack avoiding additional effort of validation and deployment
3. Replication of the azure applications as it is in cloud but on-premises
Azure Stack not only support IOT edge scenario's. Any plant level business application can also be deployed and be available to the plant irrespective of network availability.
4. Reduction in Development, Deployment and maintenance cost - single development
All the resources (VM's, SQL Server and MySQL) which are created in Azure stack are the same as it is when created in Azure Cloud. This reduces the portability issues with respect development scenarios when building first response applications. Applications developed can be deployed in Azure cloud for validation before pushing it to the remote location.
5. Enables Azure High Performance Compute products on-premises.
Most of the Azure IoT edge specific computation apps are slowly getting migrated towards Azure Container Based Apps. This enables to deploy the high performance applications like Stream Analytics, Azure ML to your local.
6. Easy Network configuration and communication between cloud and on-premises architecture
Azure Hybrid environment is a well proven across domains and various industry levels. Configuration of Azure stack on-premises and cloud could be achieved by proven ways and methodologies.
Microsoft provide or allows to you deploy the base components(Azure IoT, Stream Analytics, Azure ML Modules) in your "device" which they call as IoT edge device. The "IoT Edge Device" is not specified by the Microsoft Azure IoT team, so assumption is any device. But the real trick is the all the modules of Azure IoT Edge are built based on Azure Container Services(AKS) and not sure of portability and how extensible are they apart from Azure container services, of course AKS are basically Azure Kubernetes Services. Still Azure Container Services are built based on the Azure platform and compatibility and deployment of Azure container services in other infrastructure is still a big question.
What Microsoft Says, Why Azure Stack
Microsoft provide or allows to you deploy the base components(Azure IoT, Stream Analytics, Azure ML Modules) in your "device" which they call as IoT edge device. The "IoT Edge Device" is not specified by the Microsoft Azure IoT team, so assumption is any device. But the real trick is the all the modules of Azure IoT Edge are built based on Azure Container Services(AKS) and not sure of portability and how extensible are they apart from Azure container services, of course AKS are basically Azure Kubernetes Services. Still Azure Container Services are built based on the Azure platform and compatibility and deployment of Azure container services in other infrastructure is still a big question.
Conclusion
After doing my research for this, I feel "Cloud Industry" is maturing very well and in the right direction. When I moved to cloud 5 or 6 years before, the whole of I.T Industry's word was "Everything will be moved to cloud, nothing will be physical or local". But now slowly I feel is, realistic and mature on what are the capabilities and where the "Thin line" of differentiating is happening in the Industry and Azure Stack will play a big role in drawing that thin line.
No comments:
Post a Comment