Comparison between Alibaba Cloud Elasticsearch clusters and self-managed Elasticsearch clusters - Elasticsearch

Alibaba Cloud Elasticsearch provides a fully managed Elasticsearch service and is fully compatible with open source Elasticsearch. Alibaba Cloud Elasticsearch optimizes kernel performance and provides the commercial plug-in X-Pack free of charge. Alibaba Cloud Elasticsearch is out-of-the-box. It features high availability and auto scaling, and supports the pay-as-you-go billing method. This topic describes the comparisons between Alibaba Cloud Elasticsearch clusters and self-managed Elasticsearch clusters in terms of costs, cluster management, supported capabilities, security, and availability.

Table 1. Costs
Item	Alibaba Cloud Elasticsearch cluster	Self-managed Elasticsearch cluster hosted on an ECS instance
Resource costs	Auto scaling is supported, which allows you to change the specifications, number, disk type, and disk space of nodes.	Resources may be insufficient during peak hours and may be wasted during off-peak hours.
Network fees	You can access a cluster over an internal network free of charge by using an Elastic Compute Service (ECS) instance that resides in the same region as the cluster. You can access a cluster over the Internet free of charge by using an ECS instance that resides in a different region from the cluster. Alibaba Cloud Elasticsearch provides the Public Network Access feature, which is enabled by default. The maximum network bandwidth for access to a cluster over the Internet is 4 GB/s.	You can access a cluster over an internal network free of charge by using an ECS instance that resides in the same region as the cluster. You can access a cluster over the Internet by using an ECS instance that resides in a different region from the cluster. However, you are charged for the access. For more information about the billing standards of the access, see Public bandwidth.
Labor and time costs	Alibaba Cloud Elasticsearch, Logstash, and Kibana (ELK) are fully managed services. They are out-of-the-box and support the pay-as-you-go billing method. A visual interface is provided for cluster O&M. This reduces O&M costs.	You must purchase a machine and manually deploy a cluster. This process requires a long period of time, and the iteration of the cluster is slow. A professional Elasticsearch engineer team is required to manage resources and perform cluster O&M. This results in high labor costs.
Costs for risk reduction	Alibaba Cloud Elasticsearch guarantees 99.9% reliability and features few IT risks, and upper-level business risks are controllable.	Service reliability is not guaranteed. In this case, technical expertise and a large number of investments are required to reduce business risks.
Feature costs	All the advanced features of the commercial plug-in X-Pack are provided free of charge. The OSS-based data backup feature is provided free of charge.	You must pay USD 5,000 for the X-Pack plug-in. You must manually back up data, and you are charged for the storage used by backups.

Table 2. Capabilities
Item	Alibaba Cloud Elasticsearch cluster	Self-managed Elasticsearch cluster hosted on an ECS instance
Usability	Clusters are out-of-the-box and support auto scaling. You can modify cluster configurations based on your business requirements with one click. You can upgrade the version of a cluster with one click. The intelligent O&M system EYou is provided. It can detect the health of more than 20 items, such as clusters, nodes, and indexes. It can also diagnose and analyze exceptions.	Cluster deployment is complex, and resources must be manually adjusted. Data must be migrated before you upgrade the version of a cluster. Cluster O&M is complex. You must run commands to view the health statuses of clusters, nodes, and indexes.
Capabilities for scenario support	All the advanced features of the X-Pack plug-in are provided free of charge. Scenario-based configuration templates are used to provide appropriate parameter configurations. A natural language processing (NLP) plug-in developed by Alibaba DAMO Academy, a vector search plug-in, and a self-developed SQL plug-in are provided. These plug-ins improve the performance of clusters in search scenarios. Cold and hot data separation is supported, and an index compression plug-in is provided. This improves the performance of clusters in logging scenarios.	You must develop the capabilities or integrate the capabilities of open source Elasticsearch on your own.
Performance	An enhanced kernel is provided to improve read and write performance.	You must ensure that the cluster performance meets your requirements, which is a complex process.
Availability	Data can be automatically backed up. Data and service reliability reaches 99.9%. A self-developed throttling plug-in and the slow query isolation feature are provided to ensure cluster stability. Multi-zone cluster deployment is supported, and an active zone-redundancy architecture is provided.	You must ensure cluster availability on your own and manually back up data. Disaster recovery is difficult to implement.
Security	Clusters are accessed over virtual private clouds (VPCs) by default. X-Pack security components are provided free of charge. Field-level access control is supported. Data can be transmitted after HTTPS-based encryption and can be stored after encryption.	The security of ECS instances is ensured, but clusters have security risks. X-Pack security components must be separately purchased.