Alibaba Cloud Elasticsearch Kernel-enhanced Edition uses AliES, which is a highly tailored kernel for Alibaba Cloud Elasticsearch. This improves Elasticsearch cluster performance and stability and optimizes the costs of using Elasticsearch in multiple scenarios. If you have high requirements for the write and query performance of your Elasticsearch cluster and want to improve the stability of your Elasticsearch cluster in scenarios in which workloads fluctuate and reduce the costs of storing massive data, we recommend that you use the Kernel-enhanced Edition of Alibaba Cloud Elasticsearch.
Feature description
The Kernel-enhanced Edition is fully compatible with open source features. In addition, the Kernel-enhanced Edition comprehensively studies and optimizes items such as the hardware model, cluster architecture, and kernel engine based on rich experience in the use of large-scale data in multiple cloud scenarios.
AliES provides the read/write splitting and storage-computing separation architectures, which significantly improve write and storage performance and optimize costs. Basic enhancements are provided in the form of plug-in free of charge. The basic enhancements that are supported vary based on Elasticsearch versions and kernel versions. You can install and configure the related plug-ins to use the basic enhancements based on your business requirements.
Basic enhancements
Category | Plug-in or feature name | Description | Elasticsearch V7.16 | Elasticsearch V7.10 | Elasticsearch V6.7 |
Performance enhancement | analytic-search | This plug-in can improve query performance in logging scenarios, accelerate queries performed on the Discover page of the Kibana console, and significantly reduce the amount of time required to complete queries by working with the concurrent query feature. For more information, see Use the analytic-search plug-in. | × | Supported in Elasticsearch clusters with a kernel version of V1.7.0 or later | × |
apack | You can configure the physical replication feature of the apack plug-in at the index level. After you enable this feature for an index, incremental index data is written to the primary shard for the index and replica shards in real time. This significantly reduces the CPU overheads of the related Elasticsearch cluster and improves write performance by 60%. For more information, see Use the physical replication feature of the apack plug-in. | × | √ | Supported in Elasticsearch clusters with a kernel version of V1.2.0 or later | |
faster-bulk | This plug-in can aggregate bulk write requests in batches based on the specified maximum request size and aggregation interval, which improves the write throughput by 20%. For more information, see Use the faster-bulk plug-in. | √ | √ | √ | |
Pruning feature for time series indexes | When you query data from a time series index, you can use the pruning feature to prune the data based on a specified time range. This improves the performance of queries based on a time series field by 30%. For more information, see Use the pruning feature for a time series index. | √ | √ | √ | |
Stability improvement | aliyun-qos | This plug-in can implement index-level read or write throttling and lower the priorities of specific indexes based on specific rules to control the read and write traffic of an Elasticsearch cluster and improve cluster stability. For more information, see Use the aliyun-qos plug-in. | √ | √ | √ |
Slow query isolation | This feature can be used to track the overheads for a single query request and implement logical separation. If the overheads for the request exceed a specific threshold, the system considers the query as an anomalous query and suspends the query. This prevents exceptions caused by a single anomalous query in the cluster and improves cluster stability. For more information, see Use the slow query isolation feature. | × | √ | Supported in Elasticsearch clusters with a kernel version of V1.3.0 or later | |
Cost optimization | aliyun-codec and codec-compression | These plug-ins support compression algorithms such as brotli and zstd, can compress various types of documents in an index, and can significantly reduce the overall storage size of the index by 40%.
If you use an Elasticsearch V7.16 or V7.10 cluster, use the aliyun-codec plug-in. For more information, see Use the aliyun-codec plug-in. If you use an Elasticsearch V6.7 cluster, use the codec-compression plug-in of the beta version. For more information, see Use the codec-compression plug-in of the beta version. | √ | √ | √ |
Feature optimization | aliyun-stream | This plug-in can be used to create, modify, query, and delete time series indexes, supports downsampling queries, and allows you to execute PromQL statements to query time series data. This significantly reduces the storage and usage costs of time series indexes. For more information, see Overview of aliyun-timestream. | Supported in Elasticsearch clusters with a kernel version of V1.8.0 or later | Supported in Elasticsearch clusters with a kernel version of V1.7.0 or later | × |
aliyun-sql | This plug-in allows you to execute SQL statements to query data and provides more abundant SQL features, such as Join and extended UDFs, than open source SQL plug-ins. For more information, see Use the aliyun-sql plug-in. | × | × | √ | |
analysis-aliws | This plug-in integrates an analyzer and a tokenizer that are developed based on the natural language processing (NLP) technology of Alibaba DAMO Academy and provides more dictionaries, which can facilitate data search and analysis. For more information, see Use the analysis-aliws plug-in. | √ | √ | √ |
References
For information about the open source features supported by different Elasticsearch versions, see Version features.
For information about how to purchase an Alibaba Cloud Elasticsearch cluster, see Create an Alibaba Cloud Elasticsearch cluster.