Today more and more API are designed in the style of Microservices which is based on distributed computing. Hence system architects have to understand the fallacies of distributed computing to avoid the failure of Microservices based API.
The fallacies of distributed computing were suggested by Sun Microsystems in 1994. Until recently there are more and more discussions of these because architects are required to possess strong knowledge in distributed computing for the successful design of Microservices based systems.
Here are eight fallacies of distributed computing. 1. Network is reliable: Computer network is built by hundreds of components. Failure of any of these components may cause network failure. In fact, the network will surely fail especially at the critical moment. 2. Latency is manageable: The routing path of network traffic is changing from time to time, especially the traffic over the Internet which has a significant latency. Even if you can assure 99% of the traffic meets the benchmark, 1% failure can screw up the entire system. 3. Bandwidth is infinite: Since Microservices based APIs are running in different servers over the network, we cannot assume there is unlimited network bandwidth. When more APIs are running, more network resources will be consumed which will slow down the network as bandwidth is not infinite. 4. Network is secure: Having a good network design is far from enough to make the network secure. We cannot assume all clients are legitimate, and we need to secure every API by having authentication, authorization and data encryption. 5. Network topology does not change: Any change of the network components, such as gateways, switchers, routers, load balancer, firewall or network providers will change the network topology which will impact the network performance, scalability and availability. 6. Network has one administrator: Most of the developers think only one administrator is required to coordinate for any change about the network. In large organizations, there will be hundreds of administrators for different parts of the network. That makes communication very hard to determine the problems of infrastructure changes. 7. Transport cost is zero: Microservices based systems run in a distributed environment. To assure the performance of the systems, a high-speed network is required. Building such a high-speed network needs substantial investment. 8. Network is homogeneous: The entire network could be partitioned into several sub-networks which are built by different equipment, devices, designs and topologies. Therefore, we can never assume a network is homogenous.
Thank you for reading this article. For the details of Fallacies of Microservice Based API, please contact us. If you want to receive more information about finance and technology, please follow our LinkedIn or subscribe to the “FinTech Insights”.
Axisoft is a Top-Notch Financial Technology Provider in Hong Kong, Singapore and China. Since 1998, we have been helping financial institutions implement global banking solutions.
Copyright © 2021 Axisoft. All Rights Reserved