Why does digital currency belong to M0
Publish: 2021-04-22 04:49:02
1. Absrtact: Recently, everyone is talking about microservices. As more and more online businesses need to provide more concurrent scale-up and scale out capabilities, microservices do provide better distributed service solutions<
Yang Xu, senior solution architect of Alibaba cloud, is the chief architect of the world's largest hybrid cloud. Four years ago, he began to be the technical director of double 11 Alibaba cloud, responsible for building the world's largest hybrid cloud structure, and implementing the "double 11" e-commerce business and technology scenarios on alicloud, And ensure that the hybrid cloud can meet the shopping needs of global customers on the day of double 11<
text:
recently, everyone is talking about microservices. As more and more online businesses need to provide more concurrent scale-up and scale out capabilities, microservices do provide better distributed service solutions
microservices are no stranger. It's easy to understand microservices when you know SOA. You can regard microservices as SOA without ESB. ESB is the bus in SOA enterprise service architecture, while micro service is a decentralized distributed software architecture. I think the biggest difference lies in the original design intention:
SOA is to maximize the reusability of complex system code
while micro service is to maximize decoupling, Different business systems can even be the communication between different languages
there is no optimal architecture, only the most appropriate architecture. All the system design principles should take solving business problems as the ultimate goal, and the technical sentiment architecture divorced from the actual business will often bring a big hole to the system. The premise of all problems is to find out how much business we are facing today and what the growth trend is. Moreover, the process of solving high concurrency must be a graal process< The whole system evolution is divided into three stages:
x-axis, horizontal expansion stage. The application server is continuously expanded horizontally through load balancing server. The most important problem of horizontal expansion is to pay attention to how to keep session and session synchronization between servers, Users can't be aware when switching between different servers. The bottleneck of DB is the number of connections, IOPs and so on
the z-axis is to split the database, and the difficulty has reached a higher level. Sharding's basic idea is to divide a database into horizontal segmentation and vertical segmentation. Horizontal segmentation is relatively simple, one master and many followers, and many masters can do it. According to the needs of business, we need to pay attention to the relationship between primary keys when designing multi master segmentation, To solve the first mock exam problem, the vertical resolution is more complex. It usually involves the transformation of the logic of the architecture. It needs to introce middleware to manage the data source. When the vertical resolution is split, the tables close to each other (such as the same mole) can be split on a library or split by hash. Thus, the original database is divided into a queue which can be expanded infinitely like a matrix
Y-axis extension, and finally function decomposition, which is what we call microservice segmentation. Microservice splitting divides giant applications into groups of different services according to their functional moles. Taobao's system also experienced this process in those years. Through the wucaishi project, we split the single war package into today's buyer, seller center, trading system, etc
two or three things you need to know before introcing microservices:
1. As the cost increases, the introction of microservice architecture needs to split the original single system. After 1 to 100, the deployment of multiple services will increase the cost
2. Solving the consistency problem of distributed transactions
in the past, a single system has many advantages, and one SQL solves all the business logic, Microservices need to involve multiple system calls to complete a task. The uncertainty of the network between systems brings a lot of uncertainty to the result. For example, today's Taobao system needs to call hundreds of systems to complete a transaction order. How to ensure the reliability of the system and the final consistency of core data such as money is something we need to understand at the beginning of design, Most of them are implemented with the help of middleware
3. The logic design principle of microservices
with the continuous splitting of microservices and the iterative development of business, chaotic calls are likely to occur between systems, so the top-level design of microservices is particularly important, and architects need to understand the architecture model of microservices. The core design idea is how to layer services and reuse services. Services are distributed by layers. The upper services are packaged by the lower services. The lower services are responsible for atomic operations. The upper services are responsible for the business composition and arrangement of the lower services. We must understand the business. Micro service splitting is not a simple system composition, Once again, we must understand the business, otherwise there will be a lot of cross calls in the upper services, and the system complexity will rise exponentially. A good microservice architect must be a business architect. Based on the business strategy, microservice design trilogy, Follow the bottom-up design principle:
atomic service
first confirm the most basic business and the most dimensional atomic service. Atomic service definition is the function that everyone will maximize reuse. It needs closed-loop operation in the application. There is no branching logic across other services. It eliminates the call to other services and has its own independent data storage, As the lowest level service abstract existence, take Taobao as an example, seller data, seller data and order data belong to the most basic atomic service
service composition
in a business scenario, a function needs to span multiple atomic services to complete an action. Composition service is to abstract business logic into independent domains. Domains need to be isolated. Service composition will use multiple atomic services to complete business logic. For example, Taobao trading platform will call user, commodity, inventory and other systems
business choreography
the outermost layer is the user oriented business process. A proctized business process needs to logically choreograph the composite service to complete the final business result. The choreography service can be completely automated, and the specific SOP definition can be completed through the workflow engine, It is also significant to improve the automation process of enterprise application. For example, Taobao's "double 11" campaign realizes the logic of marketing activities by reusing the service composition
4. The complexity of operation and maintenance management is improved
microservices increase the number of applications, and link integration, testing, and deployment become new challenges. The problem solved by the previous war package needs to be completed through multi application publishing. The dependency between services at the time of publishing will lead to the unavailability of functions, and the dependency at the testing stage may make the use cases unable to run, These will be new issues to be considered and supported by platform tools. At present, Ali uses aone procts to ensure continuous integrated delivery from daily delivery to pre delivery to online delivery.
Yang Xu, senior solution architect of Alibaba cloud, is the chief architect of the world's largest hybrid cloud. Four years ago, he began to be the technical director of double 11 Alibaba cloud, responsible for building the world's largest hybrid cloud structure, and implementing the "double 11" e-commerce business and technology scenarios on alicloud, And ensure that the hybrid cloud can meet the shopping needs of global customers on the day of double 11<
text:
recently, everyone is talking about microservices. As more and more online businesses need to provide more concurrent scale-up and scale out capabilities, microservices do provide better distributed service solutions
microservices are no stranger. It's easy to understand microservices when you know SOA. You can regard microservices as SOA without ESB. ESB is the bus in SOA enterprise service architecture, while micro service is a decentralized distributed software architecture. I think the biggest difference lies in the original design intention:
SOA is to maximize the reusability of complex system code
while micro service is to maximize decoupling, Different business systems can even be the communication between different languages
there is no optimal architecture, only the most appropriate architecture. All the system design principles should take solving business problems as the ultimate goal, and the technical sentiment architecture divorced from the actual business will often bring a big hole to the system. The premise of all problems is to find out how much business we are facing today and what the growth trend is. Moreover, the process of solving high concurrency must be a graal process< The whole system evolution is divided into three stages:
x-axis, horizontal expansion stage. The application server is continuously expanded horizontally through load balancing server. The most important problem of horizontal expansion is to pay attention to how to keep session and session synchronization between servers, Users can't be aware when switching between different servers. The bottleneck of DB is the number of connections, IOPs and so on
the z-axis is to split the database, and the difficulty has reached a higher level. Sharding's basic idea is to divide a database into horizontal segmentation and vertical segmentation. Horizontal segmentation is relatively simple, one master and many followers, and many masters can do it. According to the needs of business, we need to pay attention to the relationship between primary keys when designing multi master segmentation, To solve the first mock exam problem, the vertical resolution is more complex. It usually involves the transformation of the logic of the architecture. It needs to introce middleware to manage the data source. When the vertical resolution is split, the tables close to each other (such as the same mole) can be split on a library or split by hash. Thus, the original database is divided into a queue which can be expanded infinitely like a matrix
Y-axis extension, and finally function decomposition, which is what we call microservice segmentation. Microservice splitting divides giant applications into groups of different services according to their functional moles. Taobao's system also experienced this process in those years. Through the wucaishi project, we split the single war package into today's buyer, seller center, trading system, etc
two or three things you need to know before introcing microservices:
1. As the cost increases, the introction of microservice architecture needs to split the original single system. After 1 to 100, the deployment of multiple services will increase the cost
2. Solving the consistency problem of distributed transactions
in the past, a single system has many advantages, and one SQL solves all the business logic, Microservices need to involve multiple system calls to complete a task. The uncertainty of the network between systems brings a lot of uncertainty to the result. For example, today's Taobao system needs to call hundreds of systems to complete a transaction order. How to ensure the reliability of the system and the final consistency of core data such as money is something we need to understand at the beginning of design, Most of them are implemented with the help of middleware
3. The logic design principle of microservices
with the continuous splitting of microservices and the iterative development of business, chaotic calls are likely to occur between systems, so the top-level design of microservices is particularly important, and architects need to understand the architecture model of microservices. The core design idea is how to layer services and reuse services. Services are distributed by layers. The upper services are packaged by the lower services. The lower services are responsible for atomic operations. The upper services are responsible for the business composition and arrangement of the lower services. We must understand the business. Micro service splitting is not a simple system composition, Once again, we must understand the business, otherwise there will be a lot of cross calls in the upper services, and the system complexity will rise exponentially. A good microservice architect must be a business architect. Based on the business strategy, microservice design trilogy, Follow the bottom-up design principle:
atomic service
first confirm the most basic business and the most dimensional atomic service. Atomic service definition is the function that everyone will maximize reuse. It needs closed-loop operation in the application. There is no branching logic across other services. It eliminates the call to other services and has its own independent data storage, As the lowest level service abstract existence, take Taobao as an example, seller data, seller data and order data belong to the most basic atomic service
service composition
in a business scenario, a function needs to span multiple atomic services to complete an action. Composition service is to abstract business logic into independent domains. Domains need to be isolated. Service composition will use multiple atomic services to complete business logic. For example, Taobao trading platform will call user, commodity, inventory and other systems
business choreography
the outermost layer is the user oriented business process. A proctized business process needs to logically choreograph the composite service to complete the final business result. The choreography service can be completely automated, and the specific SOP definition can be completed through the workflow engine, It is also significant to improve the automation process of enterprise application. For example, Taobao's "double 11" campaign realizes the logic of marketing activities by reusing the service composition
4. The complexity of operation and maintenance management is improved
microservices increase the number of applications, and link integration, testing, and deployment become new challenges. The problem solved by the previous war package needs to be completed through multi application publishing. The dependency between services at the time of publishing will lead to the unavailability of functions, and the dependency at the testing stage may make the use cases unable to run, These will be new issues to be considered and supported by platform tools. At present, Ali uses aone procts to ensure continuous integrated delivery from daily delivery to pre delivery to online delivery.
2. Tiktok: first, you can record the mode and then save it to the local. Then you'd better download the mobile phone editing and edit it, then click the middle and the lower plus sign to send the local video.
278
1111
anonymous user 1 hours ago.
278
1111
anonymous user 1 hours ago.
3. This question actually leads to a topic worthy of discussion: how do we see the "use" of the existing digital currency? Or how do we divide our understanding of digital currency
this division of cognition not only helps us understand digital currency, but also guides our behavior in the process of investment. So I'm going to analyze this topic in two or three articles
from the perspective of cognition, I divide all digital currencies into the following three categories:
the first category is that their value and consensus have been strongly recognized, almost indisputable, and have been transmitted from within the circle to outside the circle. This kind of currency is the least in the whole field of digital currency
this kind of currency typically includes bitcoin and Taiwan dollar. Bitcoin is recognized as "digital gold" and has repeatedly shown the characteristics of "sea calming needle" when the external political and economic environment is unstable; The profit model of platform currency is clear, especially the top three platform currencies have solid cash flow and income relying on the exchange
when talking about this kind of currency, both their consensus and value are highly recognized in the instry, and ordinary people can understand it outside the instry
when we talk about bitcoin with ordinary people, as long as we tell them that gold goes up, bitcoin goes up; When talking about platform currency with ordinary people, just tell them that platform currency is equivalent to the stock of securities companies, and they will immediately understand what we are talking about
although some people outside the instry do not agree with some consensus (for example, some people still do not agree with the value of bitcoin), this does not prevent ordinary people from understanding what we are talking about
the second type is that consensus has been reached, but the actual use value is still being proved, or the so-called "use value" can not show us the actual effect for the time being. This kind of currency is slightly more than the first kind, but also very few in the whole field of digital currency
this kind of currency typically includes Ethereum, EOS and cross chain projects in the field of smart contract
take Ethereum and EOS for example, they are highly expected by the instry. We hope to see that their smart contract function can proce killer applications and bring higher efficiency and lower cost to our daily life
in the past two years, Ethereum has developed a highly concerned defi Ecology (defi, decentralized finance, often referred to as distributed finance or decentralized Finance) in the instry
however, does this ecology have a direct relationship with the life of ordinary people or bring direct benefits? In fact, it does not exist at all. It is still a "niche ecology" in a small circle, far away from our daily life
the ecosystem developed by EOS in the past two years is far from the areas where we expect to bring value into play, and even there are not many people in the instry who pay attention to it
the concept of cross chain project is very novel when it first came out, and it is also considered by the instry to have great potential demand. But after a period of discussion, some people began to question whether this demand really exists
the instry has a consensus, or even a high consensus, on these currencies, but their current value is not satisfactory. Therefore, even if they are recognized by the instry, it is difficult for this consensus to be transmitted outside the instry, not to mention that ordinary people can understand these consensus and values
so we can't talk with ordinary people about what "defi" and "cross chain" are and what's the use
the third category is the currency whose consensus is still in the process of forming. The design objectives and scenarios of this kind of currency may not be understood or recognized by the instry, or even there are great differences in understanding. Their value is even more difficult to judge. This kind of currency accounts for the vast majority of the current market.
this division of cognition not only helps us understand digital currency, but also guides our behavior in the process of investment. So I'm going to analyze this topic in two or three articles
from the perspective of cognition, I divide all digital currencies into the following three categories:
the first category is that their value and consensus have been strongly recognized, almost indisputable, and have been transmitted from within the circle to outside the circle. This kind of currency is the least in the whole field of digital currency
this kind of currency typically includes bitcoin and Taiwan dollar. Bitcoin is recognized as "digital gold" and has repeatedly shown the characteristics of "sea calming needle" when the external political and economic environment is unstable; The profit model of platform currency is clear, especially the top three platform currencies have solid cash flow and income relying on the exchange
when talking about this kind of currency, both their consensus and value are highly recognized in the instry, and ordinary people can understand it outside the instry
when we talk about bitcoin with ordinary people, as long as we tell them that gold goes up, bitcoin goes up; When talking about platform currency with ordinary people, just tell them that platform currency is equivalent to the stock of securities companies, and they will immediately understand what we are talking about
although some people outside the instry do not agree with some consensus (for example, some people still do not agree with the value of bitcoin), this does not prevent ordinary people from understanding what we are talking about
the second type is that consensus has been reached, but the actual use value is still being proved, or the so-called "use value" can not show us the actual effect for the time being. This kind of currency is slightly more than the first kind, but also very few in the whole field of digital currency
this kind of currency typically includes Ethereum, EOS and cross chain projects in the field of smart contract
take Ethereum and EOS for example, they are highly expected by the instry. We hope to see that their smart contract function can proce killer applications and bring higher efficiency and lower cost to our daily life
in the past two years, Ethereum has developed a highly concerned defi Ecology (defi, decentralized finance, often referred to as distributed finance or decentralized Finance) in the instry
however, does this ecology have a direct relationship with the life of ordinary people or bring direct benefits? In fact, it does not exist at all. It is still a "niche ecology" in a small circle, far away from our daily life
the ecosystem developed by EOS in the past two years is far from the areas where we expect to bring value into play, and even there are not many people in the instry who pay attention to it
the concept of cross chain project is very novel when it first came out, and it is also considered by the instry to have great potential demand. But after a period of discussion, some people began to question whether this demand really exists
the instry has a consensus, or even a high consensus, on these currencies, but their current value is not satisfactory. Therefore, even if they are recognized by the instry, it is difficult for this consensus to be transmitted outside the instry, not to mention that ordinary people can understand these consensus and values
so we can't talk with ordinary people about what "defi" and "cross chain" are and what's the use
the third category is the currency whose consensus is still in the process of forming. The design objectives and scenarios of this kind of currency may not be understood or recognized by the instry, or even there are great differences in understanding. Their value is even more difficult to judge. This kind of currency accounts for the vast majority of the current market.
Hot content