The difference between distributed storage and centralized storage
1. The physical media of centralized storage are deployed centrally, while the physical media of distributed storage are distributed to different geographical locations.
2. Centralized storage has high requirements on the computer room environment and requires a large computer room space, while distributed storage requires distributed deployment of small-capacity equipment, so it has low requirements on the computer room environment.
Choice of distributed and centralized storage
The advantages and disadvantages of centralized storage are that physical media are deployed centrally; video streaming Uploading to the center has high requirements for the computer room environment, requiring a large computer room space, load-bearing, air conditioning, etc. are all issues that need to be considered.
The advantages and disadvantages of distributed storage and centralized management are that physical media are distributed to different geographical locations; video streams are uploaded nearby, and there are no requirements for backbone network bandwidth; multiple sets of low-end, small-capacity storage devices can be used Distributed deployment means lower equipment prices and maintenance costs; distributed deployment of small-capacity equipment has low requirements on the computer room environment.
Considerations of distributed storage
1. Consistency
Distributed storage systems need to use multiple servers to store data together, and with the As the number of servers increases, the probability of server failure also increases. In order to ensure that the system is still available in the event of server failure. The general approach is to divide a piece of data into multiple copies and store them in different servers. However, due to the existence of faults and parallel storage, there may be inconsistencies between multiple copies of the same data. The property of ensuring that multiple copies of data are completely consistent is called consistency here.
2. Availability
Distributed storage systems require multiple servers to work at the same time. When the number of servers increases, it is inevitable that some of them will fail. We hope that this situation will not have too big an impact on the entire system. After a part of the nodes in the system fails, the system as a whole does not affect the client's read/write requests, which is called availability.
3. Partition Fault Tolerance
Multiple servers in a distributed storage system are connected through the network. However, we cannot guarantee that the network is always smooth. Distributed systems need to have a certain degree of fault tolerance to deal with problems caused by network failures. A satisfactory situation is that a distributed storage system can still work when a network is broken into multiple parts due to failure.
Recommended tutorial: "PHP Tutorial"
The above is the detailed content of The difference between distributed storage and centralized storage. For more information, please follow other related articles on the PHP Chinese website!