8.1 Service Mesh Overview
Emerging next-generation micro-services architecture, known as the next generation of micro-services, but also one of the representatives cloud technology native technology stack.
The origin of 8.1.1 Service Mesh
From 2016 to 2018, service mesh gone through the process from scratch
8.1.2 Service Mesh definition
Grid services layer is an infrastructure for handling communications between services. Modern cloud-native applications have a reliable delivery service complex topology, responsible for implementing the requested service grid in these topologies. In practice, the grid service is typically implemented as a set of lightweight network proxy, together with the application they deploy, transparent to the application.
8.1.3 Service Mesh Detailed
- Single service call: Application examples and Service mesh proxy instances are two separate processes, communication between the remote call them, rather than the code-level method calls. The client's request will first Service Mesh proxy instance, proxy instance showed Sidecar, complete service discovery, load balancing and other basic functions, fusing, current limiting, a retry fault tolerance, routing, and authentication, authorization, encryption, and finally sends the request to the application service.
- Multiple service calls: Service mesh between all the services responsible for forwarding the request, the service is only responsible for sending and processing requests, no longer responsible for passing the specified logical request.
- A large number of service calls: When there is a large number of service system, the relationship between the performance of network service call. The relationship between service calls Sidecar form a network, which is the origin Service Mesh (service grid) name.
- Service Mesh definition of Review: Abstract: Service mesh is an abstraction layer, Wang is responsible for the completion of inter-service communication. These functions and peeled out from the application, a single communication layer is formed, and sinks to the infrastructure level. Function: Reliable transfer request. Deployment: Lightweight network proxy to sidecar model and deploy one application, remote communication between the two calls. Transparent: Service mesh is transparent to the application. Service Mesh can be deployed independently upgrade, extension, repair defects, without having to modify the application procedure.
8.2 Service Mesh Evolution
The following describes the origins Service Mesh technology, development, and evolution process step by step.
8.2.1 Case of ancient times
Processing network communication detail in application code, such as the packet sequence and flow control. Later, TCP / IP protocol stack is responsible for these functions.
Status era of micro-services 8.2.2
Service discovery, load balancing, fuse, try again.
8.2.3 invasive pain point frame
Spring Cloud and Dubbo these traditional micro-services governance framework are invasive micro-services framework.
1 pain point: high threshold.
Pain Point 2: dysfunction. Istio routing powerful than Spring Cloud.
Pain Point 3: You can not cross-language. Micro service has the advantage of cross-language.
Pain Point 4: Upgrade difficulties.
8.2.4 idea to solve the problem
Exploring 8.2.5 Proxy mode
8.2.6 Sidecar mode appears
8.2.7 The first point of service Mesh
8.2.8 The second generation of Service Mesh
8.3 Service Mesh market competition
8.3.1 Service Mesh of infancy
8.3.2 precipitous Linkerd
8.3.3 placid Envoy
8.3.4 saddled mission Istio
8.3.5 backs against the wall Buoyant
8.3.6 Other participants
Domestic developments 8.3.7 Service Mesh of
8.4 The same
8.4.1 Istio Overview
8.4.2 architecture and core components