Shulou(Shulou.com)05/31 Report--

This article introduces the knowledge of "is Service Mesh a technology?". In the operation of actual cases, many people will encounter such a dilemma. Then let the editor lead you to learn how to deal with these situations. I hope you can read it carefully and be able to achieve something!

Where did Service Mesh come from?

In the past few months, Service Mesh has been the undisputed focus of the industry. Articles about Service Mesh and the future of software architecture are highly divided around different technology vendors, but there is one thing in common about the rapid transformation of how to use API in the enterprise and what this means for the topology of our traffic.

Service API exists primarily as an edge interface that connects external developers to internal systems, and a "glue" that binds these internal systems (microservices) to the overall functionality. Therefore, the micro-service-oriented architecture will inevitably lead to an increase in internal communication in the data center. One of the inevitable results of tongguo is an increase in internal communications within the data center. Service Mesh emerged as a potential solution to address the challenges of increased traffic in the east and west by providing a different framework to deploy existing technologies.

Service Mesh is a model, not a technology

Just as microservices are a pattern rather than a specific technology, Service Mesh is a pattern. Distinguishing between the two sounds a lot more complicated than it really is. If we think about this from an object-oriented programming (OOP) perspective, a pattern describes interfaces, not implementations.

In the context of microservices, the Service Mesh deployment model can better manage east-west traffic through sidecar proxies. As we break up the system and use microservices to build new products, the topology of our traffic is also shifting from external dominance to internal traffic growth. The increase in east-west traffic in the data center is due to the fact that we replace past function calls with network calls, which means that our microservices must communicate with each other over the network. But we all know that the network can be unreliable.

By using different deployment models, Service Mesh hopes to address the challenges of increased east-west traffic. Although the processing delay of 100ms middleware is not ideal for traditional north-south traffic, it is not unacceptable, but in the micro-service architecture of east-west traffic, such delay cannot be tolerated. The reason is that increased traffic between services from east to west increases latency, which may cause delays in API when chains of 700ms requests across different services are executed and returned.

To reduce this delay, a sidecar agent running with the microservice process is introduced to remove additional jumps from the network. The Sidecar agent, corresponding to our request to execute the data plane on the path, also provides better flexibility because we no longer have a single point of failure. It is worth noting that the sidecar agent bears the cost of having a proxy instance for each instance of our microservice, which requires a small footprint to minimize resource consumption.

From a functional perspective, API management products have provided the Service Mesh that has been provided for many years. Features such as observability, network error handling, and health check are hallmarks of API management. These capabilities themselves do not constitute any novel capabilities, but as a pattern, Service Mesh introduces new ways to deploy them in our architecture.

The traditional API management scheme can no longer keep up.

Why don't most traditional API management solutions allow this new deployment option? Because they were "born in a single world".

It turns out that the API management solution built before Docker and Kubernetes is itself a whole and is not designed to work effectively in the emerging container ecosystem. The heavyweight runtime and slower performance provided by traditional API management solutions are acceptable in traditional edge API use cases, but in microservice architectures, latency increases over time by increasing traffic activity in the east-west direction. In essence, traditional API management solutions end up being too heavyweight, difficult to automate, and too slow to effectively coordinate the increasing communication inherent with microservices.

Because developers understand this, legacy API management solutions born before containers introduce what they call "micro gateways" to handle east-west traffic and avoid rewriting their existing, bloated, single gateway solutions. The problem is that although these micro-gateways are lighter, they still need legacy solutions to run with them in order to enforce policy enforcement. This means not only maintaining the original heavy dependencies in the stack, but also increasing the latency between each request. It's not hard to understand why Service Mesh feels like a whole new category, because past API management solutions can no longer support requirements.

This is the end of the content of "is Service Mesh a Technology?" Thank you for reading. If you want to know more about the industry, you can follow the website, the editor will output more high-quality practical articles for you!

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.