Which is Better? Amazon Elasticsearch Service vs. Elasticsearch Service on Elastic Cloud
Mainly we have two alternatives for Elasticsearch usage on a cloud platform. Amazon vs Elasticsearch
Firstly I am sharing a comparison info from Elasticsearch’s corporate page.
AWS Elasticsearch Service: Amazon vs. Elastic
The Elasticsearch Service on Elastic Cloud and Amazon Elasticsearch Service are not the same. With the Elasticsearch Service, you can rest easy knowing Elastic, the company behind Elasticsearch, is backing your deployment. Only Elastic has the experts, solutions, features, and product vision that your mission-critical deployments require.
The operations gap: we give you the flexibility
Our Elasticsearch Service on Elastic Cloud is the only offering that allows users to deploy custom topologies for solving different use cases. This includes hot-warm architecture with index curation for logging and time series projects, and other templates optimized for I/O, compute, and memory.
All a user has to do is simply click and drag on a slider bar and all the cluster provisioning and optimized hardware-matching magic unfolds behind the scenes.
The knowledge gap: we’ve got the experts
There is no compression algorithm for experience or expertise. We are the creators and maintainers of Elasticsearch, Kibana, Beats, and Logstash and we’ve been running our hosted Elasticsearch Service (previously known as Found) since 2012.
Amazon introduced their AWS Elasticsearch Service offering in 2015. Amazon, of course, has been offering services via AWS since 2006, but in the context of managing and supporting hosted Elasticsearch, our team has a few years head start.
No one knows the Elastic Stack better than we do. And no one hosts and manages it better than we do.
The deployment gap: we give you options
We strive to be where our users are — or might want to be. If that’s AWS, we’re there. If that’s Google Cloud Platform (GCP), Alibaba, or Azure, we’re there, too.
If you want to migrate self-managed workloads to a public cloud, we’ve made that possible. Or if you want to take our Elastic Cloud and run it on prem or in a hybrid environment, check out ECE.
And if you’d like to operationalize or scale a search project, Elastic Cloud is our family of SaaS offerings which also includes our Site Search Service and App Search Service.
None of these options are available through Amazon Elasticsearch Service.
AWS Elasticsearch Service: Amazon vs. Elastic
| Amazon Elasticsearch Service | Elasticsearch Service on Elastic Cloud | |
|---|---|---|
| Elastic Stack features | Some | All |
| Features are formerly known as X‑Pack | No | Yes Stack Security, Alerting, Monitoring, Graph, Reporting, Machine learning |
| Canvas | No | Yes (Canvas docs) |
| Kibana Spaces | No | Yes (Spaces docs) |
| Localized Kibana (Chinese and Japanese) | No | Yes (Chinese Localization docs + Kibana in Japanese) |
| Elastic Maps | No | Yes (Maps docs) |
| Elasticsearch SQL Support | Non-native SQL engine based on NLP China’s implementation. No ODBC driver, no integrated field & document-level security. | Yes (SQL docs) |
| Data rollups (API & UI) | No | Yes (Rollup docs) |
| Centralized management for Beats & Logstash | No | Yes (Logstash docs) |
| Elastic APM with distributed tracing and service maps | No | Yes (APM docs) |
| Elastic Uptime | No | Yes (Uptime docs) |
| Elastic SIEM | No | Yes (SIEM docs) |
| Frozen indices | No | Yes (Frozen indices docs) |
| Index lifecycle management | No | Yes (ILM docs) |
| Cross-cluster search | No | Yes (Cross-cluster search docs) |
| Asynchronous search | No | Yes (Asynchronous search docs) |
| Snapshot repositories management | No | Yes (Snapshot docs) |
| Data transforms | No | Yes (Transforms docs) |
| Results pinning | No | Yes (Pinned Query) |
| Vector field and vector similarity functions | No | Yes (Vector docs) |
| Flattened datatype | No | Yes (Datatype docs) Containing multiple fields within one JSON object, facilitating a flexible record structure with a large number of areas and reducing index size |
| Histogram datatype | No | Yes (Histogram datatype docs) |
| Dynamically updateable synonyms | No | Yes (Search analyzer docs) Enabling update of query time synonyms by API call without requiring a restart |
| Cumulative cardinality aggregation | No | Yes (Cardinality docs) |
| Arbitrary coordinate systems | No | Yes (Shape field type) |
| Ingest circle geo shapes | No | Yes (Circle Ingest Processor) |
| Match and geo-match enrichment | No | Yes (Match and geo-match policies for the enrich processor) |
| Elastic Stack security | Only perimeter-level security and standard IAM policies | Transport encryption Authentication Role-based access control attribute-based access control Field- and document-level security Encryption at rest(Stack security docs) |
| Alerting | Basic alerting that relies on Amazon Simple Notification System (SNS). Third-party system integrations include Chime, Slack, custom webhook, and SNS. | Fully integrated and tailor-made for Elasticsearch and Kibana.Alerts directly from within the SIEM, APM, Metrics, and Uptime apps single UI for viewing, searching, and managing all alerts notifications via third-party tools like email, PagerDuty, Slack, ServiceNow, and more unsupervised machine learning goes beyond simple rule-based alerts(Alerting docs) |
| Machine learning | Need to build and manage your models created using additional AWS services or third-party services. | Fully integrated APIs in Elasticsearch and UI in Kibana for simplifying and supercharging your ML experience.Anomaly detectionAutoML for regression, classification, and outlier detection modelsInference processor for ML models(Machine learning docs) |
| Stack monitoring | It depends on Amazon Cloudwatch, which covers a few metrics, including cluster state, node information, etc. | Feature-rich and complete monitoring product specifically designed for Elasticsearch and Kibana.Captures a broader range of metrics, including search/index rate and latency, garbage collection count and duration, thread pool bulk rejection/queue, Lucene memory breakdown, and more, with 10-second data granularity to ensure that clusters are running healthy. Robust tools to diagnose, troubleshoot, and keep your group healthy, including automatic alerts on cluster issues. (Stack monitoring docs) |
| Graph | No | Yes (Graph docs) |
| Reporting | No | Yes (Reporting docs) |
| Custom plugin support | Not supported | Supported |
| Java transport client | Not supported | Supported |
| Operational features | Some | All |
| Current Elastic Stack version | 7.4 | 7.7 |
| Dedicated VPC support | Yes | Yes |
| REST API for deployment management | Yes | Yes (API docs) |
| Secure peering connection | Yes | Yes |
| Same-day Elastic Stack version release | No | Yes |
| Upgrade Assistant UI and APIs | No | Yes |
| Hot-warm deployment template with index curation | No | Yes |
| Default snapshots | Yes Every hour | Yes Configurable Policy Defaults to 48 times per day Every 30 minutes Stored for 48 hours _source-only snapshot also available |
| The instant rollout of Stack security patches | No | Yes |
| Cross zone replication | Support for up to 3 availability zones | Support for up to 3 availability zones |
| Uptime SLA | Yes | Yes |
| Elastic Maps Service (geo-visualizations in Kibana) | It does not work out of the box | Yes |
| Kibana Lens | No | Yes |
| SOC 2 compliance | Yes | Yes |
| CSA STAR compliance | No | Yes (Security and compliance page) |
| HIPAA BAA ready | Yes | Yes (Elastic Cloud Security page) |
| ISO/IEC 27001/27017/27018 | Yes | Yes (Security and compliance page) |
| Cloud integrations | All | All |
| AWS Kinesis, CloudWatch Logs, SQS | Yes | Yes, via Functionbeat (Functionbeat docs) |
| Support | Some | All |
| Elastic technical support | No | Yes |
| SLA-based support | General level support, not specific to AWS Elasticsearch Service | Yes |