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

This article will explain in detail how to use the BigTable model in Cassandra and HBase. The content of the article is of high quality, so the editor shares it for you as a reference. I hope you will have some understanding of the relevant knowledge after reading this article.

As we all know, BigTable is the king of NoSQL database, and its paper is the cornerstone of NoSQL theory, but it is a pity that BigTable is not open source, so there is an open source version of BigTable, among which the outstanding ones include the two mentioned today: Cassandra and HBase. The following editor will explain how to use the BigTable model in Cassandra and HBase.

How to use the BigTable model in Cassandra and HBase

1.FundamentallyDistributed (distributed storage)

From the initial planning, the project is to serve large amounts of data, of course, the idea of distributed storage is also rooted in its blood. Distributed systems mainly need to consider two aspects: partitioning (partition storage, which can also be understood as Sharding) and replication (data replication, mainly copying data into multiple copies to improve availability).

2.ColumnOriented (column storage)

Unlike a normal RDBMS, a normal RDBMS is usually stored in rows, with a row of data stored continuously on a piece of disk space. Column storage is the continuous storage of each column. Therefore, it does not take up too much space to deal with the NULL field in the field. It can also support flexible and loose column definitions. That's what we usually call schema-less.

How to use the BigTable model in Cassandra and HBase

3.Sequentialwrite (write disk sequentially)

One of the features of BigTable system is its optimization of write performance. It is written by recording an operation log, and then directly writing in the memory of the data set, and then its collection conditionally or regularly flush the data to disk. The disk operations involved in recording operation logs or data flush to disk will be sequenced. Therefore, the unnecessary disk seek time caused by disk random operation is avoided.

4.Mergedread (read operation data merge)

As mentioned above, the write operation is carried out by regularly flush the data directly to disk, and each flush generates a data block, which may result in a data in multiple data blocks, and the values in these multiple versions need to be merged when reading. The bloom-filter algorithm is used to determine whether a data block contains a specified value.

5.PeriodicDataCompaction (periodic data merging)

As mentioned above, one data may exist in multiple data blocks, and if we don't deal with it, there will be more and more data blocks over time. So the BigTable system will merge the data at regular intervals. In the process of flush the data in memory directly to the disk mentioned above, flush carried out a data sorting operation before, that is to say, the data stored in the blocks in the disk are all sequential, so the sorting and merging of a pile of sequential data is actually very similar to the well-known multi-way merge sorting. Therefore, the efficiency of timing data merging is very high.

This is the end of the BigTable model on how to use Cassandra and HBase. I hope the above content can be of some help and learn more knowledge. If you think the article is good, you can share it for more people to see.

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.