Smart system & method to efficiently load balance the work load in microservice's architecture & to improve performance..
Publication Date: 2017-Feb-08
The IP.com Prior Art Database
Smart system & method to efficiently load balance the work load in microservice's architecture & to improve performance
Smart system & method to efficiently load balance the work load in microservice 's architecture & to improve performance..
Nowadays, microservice architecture is becoming popular because of it 's advantages in cloud computing systems In microservices architecture, a complex application is divided into multiple individual microservices which can do a single task cleanly and each microservices are deployed across server machines. In microservices system, there will be more request/responses flowing into/from the server because of more microservice instances doing single micro task as compared to monolithic architecture. This causes more delay in getting responses as compared to monolithic architecture because more request/response will take more network bandwidth & also consumes more resources on the machines involved. Also, there are no existing solutions to utilize the machine resources efficiently. Also, there are no existing solutions to efficiently predict the future request from the user & to keep the response ready. This article proposes a new feature to solve these problems.
Consider the following usecase scenario:
Assume a cricket website is built on microservice architecture & there are multiple microservices deployed on the server machine such as "Match_Result_Page", "Country_ABC_versus_Country_XYZ_Match_Summary_Page", "Scorecard_Page_1", "Scorecard_Page_2", "Photo_slides", "Video_highlights", "Current_Match_Scorecard_Page".. so on. These individual microservices can be deployed on a cluster of server machines & can be load balanced with the help of load balancr (web server).
Consider if the cricket match has taken place in a different country having different timezone other than India. Then there will be millions of cricket fans in India who wake up in the morning hitting this website & in turn these microservices to know the details about the cricket match. So let us assume that in between 6 AM to 8 AM IST there will be huge load
on the the server for these microservices. In existing solutions, we can have loadbalancer to balance these incoming requests for this website across server machines but still we cannot achieve quicker response time to the users because the request load is high. So, the user has to wait longer for response & in some cases, has to see Error 404 or "Request timed out error". Also, we can cache static html page but that is not the solution always.
There are other such usecase scenarios like on festival day, there will be demand for certain product on shopping websites. For example, on Deepavali festival day, there will be demand for crackers. Similarly, on Valentine's day, there will be demand for gifts.
In other usecase scenarios, the trending topics at that point of time will be searched more. For example at the time of movie release like "Air Lift", people will search about this movie as well as the history pages having details about the real stories & real characters on which this movie was b...