At the age of 35, I made a decision to face unemployment

Preface

Recently with the boss and CTO To discuss , Finally, it is decided to turn to cloud native database , After the meeting , our CTO Hold my shoulder , And I said a meaningful sentence ： I 35 Year old , The decision to switch to cloud native databases may affect my career , Or get promoted and marry Bai Fumei , Or make a mess and eventually lose your job . I heard this sentence a little meaningful , Although our company does not fully embrace cloud native , But at present, most of the company's products have turned to the embrace of cloud native database . I am currently working with databases close to 10 Years. , front 5 In, I came into contact with traditional databases , after 5 In, with the change of technology, I began to contact cloud database , Then the company began to accept cloud native databases , There are still many feelings .

front 5 year , The pain of traditional databases

I just came out to work. That would be 12 year , At that time, mobile Internet began to rise , More and more Internet companies are beginning to rise , Data gradually has a certain scale , I got a job as a background developer , In addition to writing every day CRUD Beyond the code , You also need to maintain and manage the company's computer room , Yes, you heard it right. It's the computer room , At that time, the database was basically built by itself , Use switches and servers , In a ’ Gloomy ’ Room , I wear thick antifreeze clothes , Alone in the cold black window, tapping orders , Fortunately, there is still a light , Not too bleak .

As an amateur computer room manager , There are too many headaches , First of all, equipment procurement is a big problem , Capitalists prefer to buy two commodities with one share of money , So the configuration of the server in the early stage 、 It's a headache to finalize the switch , If the configuration is low , In the later stage, I need to spend a lot of effort to optimize and configure , I want to carry a bucket , I am a CRUD The programmer , No Linux The programmer . If the configuration is high , I will definitely relax a lot , You can focus on the preparation of business and the optimization of the overall architecture , But put , The cost is too high , Directly be pass 了 .
The second is the installation and maintenance of various software in the computer room , Because our company didn't hire special operation and maintenance , At that time, the operation and maintenance work was not very mature , There is no special operation and maintenance , Programmers do it part-time , At that time, the company had a computer room watch , good heavens , I have a look at , All are background development programmers . We are responsible for software installation, virtualization and later maintenance , pain ！ It's too painful . I drew a sketch , Maybe I have to manage and maintain so many machines a day alone .

There is also a very important point , That is the question of tariff , Because our company is the website of real estate middlemen , At that time, the business was seasonal , In about 6-12 Month is the peak season , At that time, the demand of home buyers was relatively large , Due to the increase of flow , The requirements for servers are also relatively large , At this time, the servers purchased in the first half of the year are completely insufficient , In this half year, our team basically won't write business , Basically, it's all kinds of optimization , Reduce the load on the server , I was thinking , If only there were a server that could be more flexible .

after 3 year , The light of cloud database

It's really unbearable behind , I chose the bucket . More than five years passed in the blink of an eye , At that time, the database has become increasingly mature ,MySQL Became the overlord , Percent of the market 80 All companies are using , And there are non relational databases , image redis、ElasticSearch Wait for this kind of database , Because of the increase in the amount of data , Database is becoming more and more important in the whole business .
At this time, cloud database appears , Our company is a small enterprise in rapid development , Using Amazon cloud native database , I was amazed by this mode of use the first time I used it . It's really convenient , I don't need to worry about the database at all , Because they all configured it for me , I just need to focus on my sql Just write it , Once there are a lot of requirements , I'll go through the approval and add money to upgrade the capacity , Once it's off-season, I can reduce the capacity , I even experienced a serverless version Aurora, No, it doesn't cost much , It saved a lot of money .

And even the money of the traditional computer room is saved , It is said that after all the products of this enterprise turned to cloud native database , Transform the original computer room of the company into a gym .

Just these years , Terminal tools are also gradually increasing , image XSehll、MobaXterm Tools such as , You can connect to the cloud directly through this tool , I even miss the days when I used to debug the database in Xiaohei air-conditioned room .

It's convenient not to say , And the variety is really complete , If you want a relational database , You can choose Amazon RDS、Amazon Aurora, If you want to choose cache , You can choose Amazon Elasticache、Amazon MemoryDB for Redis, Want to use a document database , Yes Amazon DocumentDB, Want to try graph database , You can choose Amazon Neptune, Even non relational databases , Yes Amazon DynamoDB, It can be said that only you can't use , You can't do it without Amazon cloud technology .

this 2 year , Volume of cloud native database

These two years , I have also gradually stepped into the pace of cloud primordial , It's not just because major manufacturers have begun to release their own cloud native products , More and more rolling , what's more , With the popularity of artificial intelligence and big data , Cloud primordial is the real future . The cloud database we used before actually belongs to Cloud Computing , Cloud computing applications are often developed internally using traditional infrastructure , And after adjustment, it can be accessed remotely in the cloud , The simple understanding is that I link the cloud software of the atmosphere through remote software on the ground , In essence, it is partly on the cloud . Next, let me talk about why I want to be completely on the cloud ？

First of all, from the perspective of use , Cloud computing applications often require manual upgrades , For example, I want to upgrade the database or server , Usually, you need to close it first, then upgrade it, and then open it , This can cause the application to break and shut down , When I use cloud native database, I feel the friendliness of cloud native database , It is highly scalable , The cluster size can be adjusted in real time , It doesn't interfere with the entire application . It means that I can upgrade or downgrade without feeling .
Then I feel more about the price , After all, we go to the cloud to reduce costs , From the previous self built computer room to the cloud database behind , Are further reducing costs , Or it is to pass the cost on to other manufacturers , We build the machine room 、 Server running 、 Server maintenance 、 A series of costs such as software installation are passed on to a number of manufacturers such as Amazon cloud technology , We only need to pay a small part of the money to enjoy better services .
Amazon cloud technology's cloud native applications do not require any hardware or software investment , Because they're done on the cloud , You can flexibly use the elastic advantages of cloud , So it's relatively cheap to use , This is a blessing for small and medium-sized enterprises .

Private goods ：Amazon DocumentDB Using experience of

By the way, let's talk about the most frequently used product of the company's transformation to cloud native database , That's it Amazon DocumentDB. It is a compatible MongoDB Database products , Due to some business development done by the company during the transformation , For massive json Data processing has performance and cost considerations .
We MongoDB It is a document database , He used json Formatted data , Different from the traditional relational database . In the past , If you want to support highly concurrent requests , Usually I need to set up multiple servers to form a cluster , and Amazon DocumentDB By independently expanding computing and storage , Support read requests of millions of documents per second , Just ask Lao tie , This function is six or six .
Our company's comment is to use MongoDB To store the , Deploy as before , Once there is a problem with the server , It will not only cause problems with the comment function , Carelessness will also lead to the downtime of the whole system , and Amazon DocumentDB A function of has helped me a lot , That is automatic hardware presetting 、 repair , I haven't worried about stability anymore , And it can also be copied automatically 、 Continuous backup and strict network isolation 99.999999999% The durability of , The stability guarantee of the data reassures me .
from MongoDB Migrate to Amazon DocumentDB It's also very simple. , With another service Amazon DMS That's all right. .

in general , This round of experience is not bad , At present, the performance is better than MongoDB The performance of cluster deployment should be a little higher , Probably saved a lot of money for the boss .

summary

35 year CTO My feeling is too deep , Recall the technology I just started using and the technology I now use , It's very different , The demand is becoming more and more difficult , And we are constantly improving , Believe in the future , Better cloud native data can be used more widely , Instead of just staying at the developer level .