|
1 | 1 | // Module included in the following assemblies:
|
2 | 2 | //
|
3 |
| -// * rosa_hcp/rosa-hcp-sts-creating-a-cluster-quickly.adoc |
| 3 | +// * rosa-architecture-models.adoc |
4 | 4 |
|
5 | 5 | :_mod-docs-content-type: CONCEPT
|
6 | 6 | [id="rosa-hcp-classic-comparison_{context}"]
|
7 |
| -= Comparing ROSA with hosted control planes and ROSA Classic |
8 |
| - |
9 |
| -{hcp-title-first} offers a different way to create a managed {product-title} (ROSA) cluster. {hcp-title} offers a reduced-cost solution with focuses on reliability and efficiency. With a focus on efficiency, you can quickly create a new cluster and deploy applications in minutes. |
10 |
| - |
11 |
| -{hcp-title} requires only a minimum of two nodes making it ideal for smaller projects while still being able to scale to support larger projects and enterprises. |
| 7 | += Comparing ROSA with HCP and ROSA Classic |
12 | 8 |
|
13 | 9 | .ROSA architectures comparison table
|
14 | 10 |
|
15 | 11 | [cols="3a,8a,8a",options="header"]
|
16 | 12 | |===
|
17 | 13 | | {nbsp} +
|
18 |
| -| Hosted Control Plane |
19 |
| -| Classic |
20 |
| - |
21 |
| -| *Cluster infrastructure hosting* |
22 |
| -| {hcp-title} deploys control plane components, such as etcd, API server, and oauth, that are hosted separately on AWS in a Red Hat-owned and managed account. |
23 |
| -| ROSA Classic deploys the control plane components side by side with infrastructure and worker nodes that are hosted together in the customer’s same AWS account. |
24 |
| - |
25 |
| -| *Provisioning Time* |
26 |
| -| Approximately 10 minutes |
27 |
| -| Approximately 40 minutes |
| 14 | +| *Hosted Control Plane (HCP)* |
| 15 | +| *Classic* |
28 | 16 |
|
29 |
| -| *Architecture* |
30 |
| -| |
31 |
| - * Underlying control plane infrastructure is fully managed |
32 |
| - * Customer can access control plane infrastructure through dedicated and explicitly exposed endpoints |
33 |
| - * Worker nodes are hosted in the customer's AWS account |
34 |
| -| |
35 |
| - * Customer is responsible for hosting control plane and AWS infrastructure, while still being _managed_ by Red Hat |
36 |
| - * Worker nodes are hosted in the customer's AWS account |
| 17 | +| *Control plane hosting* |
| 18 | +| Control plane components, such as the API server etcd database, are hosted in a Red Hat-owned AWS account. |
| 19 | +| Control plane components, such as the API server etcd database, are hosted in a customer-owned AWS account. |
37 | 20 |
|
38 |
| -| *Minimum Amazon EC2 footprint* |
39 |
| -| One cluster requires a minimum of two nodes |
40 |
| -| One cluster requires a minimum of seven nodes |
| 21 | +| *Virtual Private Cloud (VPC)* |
| 22 | +| Worker nodes communicate with the control plane over link:https://docs.aws.amazon.com/vpc/latest/privatelink/what-is-privatelink.html[AWS PrivateLink]. |
| 23 | +| Worker nodes and control plane nodes are deployed in the customer's VPC. |
41 | 24 |
|
42 |
| -| *Deployment* |
43 |
| -| |
44 |
| - * Deploy using the ROSA CLI (`rosa`) |
45 |
| - * Customers provision "Hosted Clusters" that deploy the control plane components into Red Hat's AWS account |
46 |
| - * Customers provision "Machine Pools" that deploy worker nodes into the customer's AWS account |
47 |
| -| |
48 |
| - * Deploy using the ROSA CLI or the web UI |
49 |
| - * Full cluster provisioning occurs in customer's AWS account |
50 |
| - |
51 |
| -| *Upgrades* |
52 |
| -| Selectively upgrade control plane and machine pools separately |
53 |
| -| Entire cluster is upgraded at one time |
54 |
| - |
55 |
| -| *Regional Availability* |
56 |
| -| |
57 |
| -* US East - N. Virginia (us-east-1) |
58 |
| -* US East - Ohio (us-east-2) |
59 |
| -* US West - Oregon (us-west-2) |
60 |
| -* Africa - Cape Town (af-south-1) |
61 |
| -* Asia Pacific - Hyderabad (ap-south-2) |
62 |
| -* Asia Pacific - Jakarta (ap-southeast-3) |
63 |
| -* Asia Pacific - Melbourne (ap-southeast-4) |
64 |
| -* Asia Pacific - Mumbai (ap-south-1) |
65 |
| -* Asia Pacific - Seoul (ap-northeast-2) |
66 |
| -* Asia Pacific - Singapore (ap-southeast-1) |
67 |
| -* Asia Pacific - Sydney (ap-southeast-2) |
68 |
| -* Asia Pacific - Tokyo (ap-northeast-1) |
69 |
| -* Canada - Central (ca-central-1) |
70 |
| -* Europe - Frankfurt (eu-central-1) |
71 |
| -* Europe - Ireland (eu-west-1) |
72 |
| -* Europe - London (eu-west-2) |
73 |
| -* Europe - Milan (eu-south-1) |
74 |
| -* Europe - Stockholm (eu-north-1) |
75 |
| -* Middle East - Bahrain (me-south-1) |
76 |
| -| For AWS Region availability, see link:https://docs.aws.amazon.com/general/latest/gr/rosa.html[Red Hat OpenShift Service on AWS endpoints and quotas] in the AWS documentation. |
77 |
| - |
78 |
| -| *Compliance* |
79 |
| -| |
80 |
| - * Compliance certifications and FIPS are not yet available. |
81 |
| -| |
82 |
| - * Compliance specifics are located in the {product-title} documentation. |
83 |
| -|=== |
| 25 | +| *Multi-zone deployment* |
| 26 | +| The control plane is always deployed across multiple availability zones (AZs). |
| 27 | +| The control plane can be deployed within a single AZ or across multiple AZs. |
84 | 28 |
|
85 |
| -[id="rosa-hcp-classic-comparison-networks_{context}"] |
86 |
| -== ROSA architecture network comparisons |
| 29 | +| *Machine pools* |
| 30 | +| Each machine pool is deployed in a single AZ (private subnet). |
| 31 | +| Machine pools can be deployed in single AZ or across multiple AZs. |
87 | 32 |
|
88 |
| -ROSA Classic and ROSA with HCP offer options to install your cluster on public and private networks. The following images show the differences between these options. |
| 33 | +| *Infrastructure Nodes* |
| 34 | +| Does not use any dedicated nodes to host platform components, such as ingress and image registry. |
| 35 | +| Uses 2 (single-AZ) or 3 (multi-AZ) dedicated nodes to host platform components. |
89 | 36 |
|
90 |
| -.ROSA Classic deployed on public and private networks |
91 |
| -image::156_OpenShift_ROSA_Arch_0621_private_public_classic.png[ROSA deployed on public and private networks] |
| 37 | +| *OpenShift Capabilities* |
| 38 | +| Platform monitoring, image registry, and the ingress controller are deployed in the worker nodes. |
| 39 | +| Platform monitoring, image registry, and the ingress controller are deployed in the dedicated infrastructure nodes. |
92 | 40 |
|
93 |
| -.ROSA with HCP deployed on a public network |
94 |
| -image::ROSA-HCP-and-ROSA-Classic-public.png[ROSA with HCP deployed on a public network] |
| 41 | +| *Cluster upgrades* |
| 42 | +| The control plane and each machine pool can be upgraded separately. |
| 43 | +| The entire cluster must be upgraded at the same time. |
95 | 44 |
|
96 |
| -.ROSA with HCP deployed on a private network |
97 |
| -image::ROSA-HCP-and-ROSA-Classic-private.png[ROSA with HCP deployed on a private network] |
| 45 | +| *Minimum EC2 footprint* |
| 46 | +| 2 EC2 instances are needed to create a cluster. |
| 47 | +| 7 (single-AZ) or 9 (multi-AZ) EC2 instances are needed to create a cluster. |
| 48 | +|=== |
0 commit comments