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

In this issue, the editor will bring you how to explore the cloud integration of Swift on the GAIA platform. The article is rich in content and analyzes and narrates it from a professional point of view. I hope you can get something after reading this article.

In stage S1, when using SwiftUI to write SOT APP used within the group, it was fortunate to participate in the exploration of cloud integration of GAIA (FaaS) platform, realized a set of Runtime Framework that complied with GAIA Funtion standard based on Swift language from beginning to end, and completed the exploration of integrated link from client to back end using a unified language stack.

As a pure iOS Native developer, for the technical sense of the back-end, most of them left behind the forum management system that I did at school, and FaaS Serverless and so on are some of the more cutting-edge technical points in the back-end field, especially the Swift language, which is still a newborn calf in the back-end, has gone through numerous detours during this period, but also learned a lot of new knowledge.

The following is the phased summary and thinking of Swift On GAIA. As there is more cross-end knowledge in this technology exploration, the understanding from the perspective of a mobile engineer may be very one-sided and wrong. If readers find that there is a misinterpretation of the concept, you are welcome to correct it in the message area. As there has been a lot of exploration in the front-end ecology on the technology stack, and the exploration and technology reserve on the Native side lags behind and the front-end, some implementations will change with the exploration of cloud integration, which is not a complete solution. You are welcome to actively discuss it for the benefit of the ecology.

Conceptual introduction

Serverless

The beginning of Serverless is an early term, as early as 2012, when I just picked up my small schoolbag and walked into the university, but the early theoretical basis has been put forward.

With the emergence of AWS's Lambda products in 2014, Serverless provided a new architecture for cloud applications. Serverless began to catch fire, and then major cloud computing vendors joined. Google Cloud Functions, Azure Funcions, IBM OpenWhisk Aliyun and other domestic cloud computing manufacturers, such as Huawei Cloud, Tencent Cloud, Baidu Cloud, Serverless products have sprung up everywhere in just a few years.

With the rapid development of container technology, IoT,5G, technology, the technical demand for decentralization, lightweight virtualization, fine-grained computing and other technologies is increasingly strong, while Serverless will take advantage of the opportunity to develop rapidly, the future rich client research and development model changes, we technicians need to continue to explore and create.

In the process of understanding Serverless, let's take a look at the evolution history of cloud services. Cloud computing has experienced physical computer room-> IaaS-> PaaS-> SaaS-> FaaS/BaaS.

▐ IaaS

IaaS (Infrastructure as a Service) Infrastructure as a Service. Refers to the provision of IT infrastructure as a service through the network. In this service model, users do not have to build a data center themselves, but use infrastructure services, including servers, storage and networks, by leasing. At this level, companies usually buy storage, network basic services.

▐ PaaS

PaaS (Platform as a Service) platform is a service. Service providers provide infrastructure services, operating system (Windows,Linux), database server, Web server, load balancer and other middleware. Relative to Iaas customers, they only need to control the upper application deployment and application hosting environment. Usually at this level, users usually buy an operating system.

▐ SaaS

SaaS (Software as a Service) Software as a Service, SaaS is a mode of providing software through Internet. Users no longer need to buy software, but rent Web-based software from providers to manage business activities without maintaining the software. Service providers will have full authority to manage and maintain the software, usually these commonly used software have databases, network services, and so on.

You can feel it visually through a picture of the Microsoft Azure service.

The GAIA platform greatly shortens the back-end R & D link, increases the delivery cost, and shields the details of operation and maintenance, which can be easily understood by a developer across the stack.

GAIA Container Architecture

Having seen the concept of GAIA, it is necessary to have a simple understanding of the GAIA container architecture.

As an end-to-end developer, it coincides with the use of SwiftUI to build SOT APP mobile version, not only to meet the needs, but also to explore new technologies.

But as a data market of APP, where does the data come from? From the backend database, since there has never been a mobile APP before, the backend students do not provide an MTOP interface for mobile use. They learn that the existing HSF service can be invoked on the FaaS platform and the model in the domain can be returned. The GAIA platform can be published as a MTOP interface, which just meets the needs.

In the previous research, it is found that the platform language stack of GAIA chooses the back-end ace language Java,Function Runtime Framework to support only Java. Later, with the exploration of the front-end stack and the Flutter exploration of Xianyu team, the platform already supports Dart Runtime and Node.js Runtime.

As an iOS Native developer, what if you don't have a familiar language? Swift announced the release of 5.x version of WWDC in 2019, and the Swift5.x version of ABI is stable, which marks the official entry of Swift into the ranks of mature languages, coupled with the fact that Swift was born as a cross-platform language, and open source Server side workgroup promotes the exploration of Swift in the back-end field, and some well-known libraries Kitura Vapor Perfect Zewo have emerged.

So let's try to build a set of Function Runtime Framework with Swift. Let's first see what Runtime Framework is about.

Swift Runtime Framework needs to do the following

Initialized by the Swift runtime.

Log monitoring

Communication protocol

The group already has middleware calls

Function scheduling

Release system construction

Services deployed in the backend are flexible, and GAIA is no exception. Swift Function Runtime and Swift Function delivery are all made into Docker Image unified delivery. Simplify as shown in the figure.

Officially go online

After Funtion Runtime Framework is completed, you can create functions through GAIA's publishing platform and write your own business code. Instead of sharing how to do this, you can go directly to the sample code and the effect diagram.

/ / File.swift/ Created by nero on 2019/9/12.//import GAIAimport HSFimport Foundationstruct EventResult: GAIA.EventResult {var isSuccess: Bool = true var code: String = "200" var message: String = "SOT Test" var payLoad: TestContent init (payLoad: String) {self.payLoad = TestContent (content: payLoad)}} struct MtopUserQuery: Codable {let pageNo: String let pageSize: String let appName: String} struct PageQuery: HSFModel {let javaClassName = "com.xx.PageQuery" let request: PageQueryRequest let pageNo: Int let pageSize: Int init (pageNo: Int PageSize: Int AppName: String) {self.pageNo = pageNo self.pageSize = pageSize self.request = PageQueryRequest (appName: appName)} let apiConsumer = PageQueryAPIConsumer ()} struct PageQueryAPIConsumer: HSFModel {let javaClassName = "com.xx.ApiConsumer" let clientName = "SOT-APP" let token = "xxx"} struct PageQueryRequest: HSFModel {let javaClassName = "com.xx.AppQueryRequest" let appName: String} class Handler: GAIA.FunctionHandler {init () {} var consume: ConsumerService func handle (when event: Event Context: FunctionContext) throws-> AnyEventResult {switch event.payLoad {case .mtop (body: let mtopRequest): do {let query = try JSONDecoder () .decode (MtopUserQuery.self, from: mtopRequest.data) let sotQuery = PageQuery (pageNo: query.pageNo, pageSize: query.pageSize, appName: query.appName) let jsonStr = try consume.invoke (methodName: "queryApp" Args: [sotQuery]) return EventResult (payLoad: jsonStr). ErasedEventResult} catch {} default: break } return EmptyEventResult () .erasedEventResult} func active (when context: FunctionContext) {} func destroy (when context: FunctionContext) {} func isHealth (when context: FunctionContext)-> Bool {return true}}

Thoughts on the Integration of GAIA Cloud

The first phase of Swift On GAIA's exploration of cloud integration has come to a successful conclusion, and an iOSer can also develop front-end and front-end programs at the same time, which also creates some new possibilities for the R & D model. Recently, I have heard many times what the bosses have shared, such as Xianyu's team's exploration of the cloud-integrated R & D model using Dart+flutter. There are some immature perspectives, please record for a while.

What are the difficulties in daily R & D and debugging of the technology stack?

In the experience of R & D mode exchange and some practical exploration, it is found that the R & D style on the end is quite different from that of the back-end, and the students on the end tend to interrupt debugging and preview the effect in real time. The force behind is that the end needs faster delivery and the end environment is easier to build, but the back-end students are more inclined to log system and link tracking. But in fact, whether it is breakpoint debugging or logging system, essentially need an effective troubleshooting system, otherwise in the actual development, because the cross-technology stack will increase the cost of debugging, once the system becomes complex, it will slow down the efficiency.

Upgrading of engineering structure

In the actual experience of Swift On GAIA, the development tools and application delivery methods used are not unified, and each language stack has its own R & D style, tool chain support, R & D model and tool chain system also need to be upgraded to avoid long links and tool chain fragmentation.

What is the boundary of the domain?

At present, Xianyu's boss has completed the support of the Dart Function environment in GAIA, which allows client students to go further, complete the back-end ability by themselves, achieve a closed-loop business without upstream and downstream dependence, and shield the details of the back-end.

However, in the analysis of some details, it is found that different business scenarios encounter different challenges. If a Function only uses the existing infrastructure, such as interface aggregation, simple logical processing is more appropriate, but if the business even the producer of the data is written by the client students, the complexity will become greater, because the cross-end way of thinking is different, the big front-end students may not think about more storage design. It may not be easy to expand as the business expands in the future.

So in the implementation of the business, how to define the boundaries of some areas is a point worth thinking about, we need some methodology or guiding principles to help Function students decide the scalability of this technology selection.

Upgrade the organizational structure of personnel

If the R & D model is greatly upgraded and the traditional back-end API is released, the relationship between upstream and downstream personnel connected to each end is bound to change greatly, and how to find the way of personnel organization also needs to be redefined and adjusted. The distribution of work is different from the younger, and the organizational structure needs to adapt to the times when it is bigger.

Middleware language neutrality, platform neutrality

At present, most of the back-end language stacks are Java, in which some group teams use CPP, and several CPP versions of SDK have emerged to call Java's ecological framework, such as the HSF CPP version of Tail CPP. With the increase of GAIA platform access to language stacks and rich client Swift/Kotlin/Java/Javascript, do you want to do something like Protobuffer to solve language differences through neutral DSL definitions?

Here, the author does not agree with a set of language to solve all environments. First, the current ecology of hand Amoy is divided into Native+Weex+H5+ Mini Program, which itself is difficult to aggregate. Second, with iOS alone, cross-platform under Apple restrictions always has some trade-offs, which can not solve all the problems.

Domain specific / generic

In the future, whether the Swift will move forward to solve the general capability on the FaaS platform or build a domain model, the construction of domain features can build a domain ecological API, define the business model, similar to the style of star rings, if the construction is a general capability, the solution may be another.

This is what the editor shares with you about Swift's exploration of cloud integration on the GAIA platform. If you happen to have similar doubts, please refer to the above analysis to understand. If you want to know more about it, you are welcome to follow the industry information channel.

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.