Estimating the cost of migrating servers to the public cloud
Use the cloud migration simulation to evaluate the opportunity and cost effectiveness of migrating servers (compute resources or VMs) that are hosted in your on-premises infrastructure to a public cloud. The simulation results enable you to quickly determine the projected cost of using the public cloud and compare the cost with that of your on-premises infrastructure.
You can simulate the migration of selected servers (from the Servers page). For every server to be migrated, you can view the optimized recommendations for instance sizes and other associated characteristics on the public cloud. You can modify these characteristics and evaluate the impact of the change on the overall migration cost.
You can simulate migration to the following public cloud providers:
- AWS (PAY-AS-YOU-GO and AWS Government subscriptions)
- Azure (PAY-AS-YOU-GO and Azure Government subscriptions)
- Google Cloud
The backend simulation algorithm in the Cost Estimation service uses one of the following strategies:
- Utilization-based migration: Resource utilization of the servers to be migrated is measured and analyzed to suggest the best or ideal resized instance types on the public clouds. This is the default strategy.
- Lift-and-Shift: Servers are replicated or re-hosted on the target public cloud without any redesigning.
For more information about the strategies, see Strategies-for-cloud-migration-simulation.
The Cost Estimation service considers a month of 30 days to estimate the migration costs.
The cloud migration simulation feature enables you to achieve the following goals:
- Compare the on-premises and public cloud costs for the same set of servers.
- Evaluate the estimated monthly cost of migrating to the public cloud.
- Evaluate the characteristics of the proposed or recommended instances on the public cloud.
- Customize or modify the characteristics of the proposed instances and evaluate the impact on the migration cost.
Prerequisites
Ensure that the administrator has completed the following tasks:
- Installed Cloud Cost Control
- Granted you necessary access
- Configured the required migration simulation settings
To estimate the cost of migrating servers to the public clouds
1. Initiate the migration simulation
You can simulate migration of selected servers from the Servers page. By default, you can migrate 1000 servers in one attempt. If the number of servers to be migrated exceeds this limit, only the first 1000 servers are migrated.
- In the navigation pane of the Helix Capacity Optimization Dashboard, select Cloud Cost Control and then, click the Servers tab.
- Click Simulate Cloud Migration.
From the list of the servers, select the servers that you want to migrate.
- Click Apply.
The Simulate migration for # servers page opens, where # is the number of servers that you selected in the earlier step.
By default, the cloud provider with the lowest projected cost of migrated servers is selected.
2. Evaluate the simulation results
Depending on the cloud provider that you want to evaluate further, click the target cloud provider name.
The page displays a table that lists the servers to be migrated. The table has two sections: Before and After. Values of the selected server's characteristics are displayed in the Before section and those of the proposed instance on the public cloud are displayed in the After section.Review the total migration cost estimates and characteristics of the proposed migrated instances on the public cloud.
The following screenshot shows the Simulation results page when simulation is initiated from the Business Services page. The Azure cloud provider is selected as the target cloud provider to display the simulation details.Label
Description
A
Estimate of the total monthly cost of the servers that you want to migrate.
B
Estimate of the total monthly cost of migrating to the public cloud providers. Costs for the following payment modes or billing options are provided (Left to right):
- On-demand or pay-as-you-go
- Full upfront yearly payment1
- Full upfront payment for three years 1
1 - For AWS, the costs are calculated based on the AWS EC2 Reserved Instances Pricing. For Azure, the costs are calculated by considering a discount of 20% on the total monthly cost because Azure does not publish a catalog of prices for reserved instances. For Google Cloud, the costs are calculated based on the committed use discounts for the predefined machine types (custom machine types are not supported).
The simulation provides an estimate of the migration costs based on the characteristics of the selected servers. Your actual monthly costs will be based on your usage, and might vary from these estimates.
If a provider does not have an instance type that matches the required configuration of a source server after simulating the migration, the next best available instance type is proposed.
A warning symbol is displayed next to this Provider as shown in the following screenshot.
C
Net percentage increase or decrease in cost after migration. This estimated value indicates whether the migration is cost-effective or not.
D
Suggested region where the migrated server will reside. You can modify the region and view the impact of this change on the cost of migration. Select a region that is closer to your user base.
E
Configuration details of the server to be migrated are displayed in the Before section. Values for the following characteristics are displayed:
Instance type, operating system that the instance is running, estimated cost of the server, number of CPU cores, allocated memory, and amount of storage or storage size.
The three bar graphs depict the estimated utilization of the CPU, memory, and storage space in the last 30 days.
If the Utilization-based strategy is applied, resource utilization is computed by considering the selected Optimization behavior. For more information, see Strategies-for-cloud-migration-simulation.
If the Lift-and-shift strategy is applied, the computation of the resource utilization considers the average value of the hourly samples. Then, 95th percentile of the hourly value over the last 30 days is computed for each resource. Spikes in the resource utilization within the hour are not considered.The utilization bar colors are based on the following default threshold ranges:
Color
Resource utilization threshold range
CPU
Memory
Storage
Green
Below 75%
Below 80%
Below 85%
Yellow
Between 75% and 85%
Between 80% and 90%
Between 85% and 95%
Red
Above 85%
Above 90%
Above 95%
You can modify the thresholds from the Metric thresholds page. For more information, see Configuring-and-managing-thresholds-for-metrics-and-indicators.
F
Configuration details of the proposed server or instance on the public cloud are displayed in the After section. Values for the following characteristics are displayed:
Instance type, operating system that the instance is running, estimated cost of the server, number of CPU cores, allocated memory, amount of storage or storage size, and storage type.
Information: The default storage type of the proposed server is as follows:
- AWS - General Purpose
- Azure - Unmanaged Standard - LRS
- Google Cloud - Local SSD
The three bar graphs depict the estimated future utilization of CPU, memory, and storage space of the target server.
The utilization values are computed by considering the estimated utilization of resources of the source server (as shown in the bar graphs in the Before section) and the amount of resources of the target server.
If Utilization-based strategy is applied, computation of the CPU utilization value also considers CPU benchmarks. For more information about this strategy, see Strategies-for-cloud-migration-simulation.
You can modify some of the characteristics of the proposed server. For more information, go to the next step.
G
The selected migration strategy and optimization behavior that are used by the backend simulation algorithm.
For more information about the supported strategies and the optimization behaviors, see Strategies-for-cloud-migration-simulation.
You can modify the selected strategy and behavior from the Settings page. For more information, see Configuring-migration-simulation-and-recommendation-settings.
Review additional details of the proposed server.
Click the Editicon that is placed next to the corresponding source server.
The Simulation migration for <source server> window opens. Values of the characteristics of the source server and proposed server that are shown on the Simulation results page are displayed in this window along with additional information.
The following additional information is displayed for the source and target servers in the Before and After sections, respectively:Additional characteristic
Description
Instance Size
Configuration of the instance or server. For example, 1 CPU, 2 GB Mem.
Estimated SPECint_rate2006
Estimated benchmark for the CPU. SPECint_rate2006 is used as the reference benchmark.
For the target server, the information in brackets indicates whether its CPU performance is faster or slower than the source server.
If the CPU benchmark value for the source server is unavailable, the simulation results are generated without considering benchmarks.
Accuracy icon
Displayed at the top-right corner in the After section. The icon indicates the accuracy level of the recommended or proposed resized configuration of the target server.
For Utilization-based migration strategy, the accuracy level is based on the Optimization behavior, defines the backend process of computing the utilization of the source server, which impacts the resizing suggestion.
3.Customize the proposed server characteristics as per your requirement and re-evaluate the migration costs and other updated values
- Ensure that the Simulation migration for <source server> window is open that displays the details of the proposed server that you want to modify.
To open the window, for the proposed server that you want to customize, click the Editicon that is placed next to the corresponding source server. The Simulation migration for <source server> window opens.
Based on your requirement, in the After section, modify the following characteristics:
Characteristic/Field
Description
Instance Type
Modify the instance type based on the catalog of instances that is available for the selected public cloud provider. You can decide whether you want to modify the instance type by reviewing the other characteristics of the proposed server. Changes to the instance type changes the Cost value for the server and causes a change to the vCPU, memory, and Storage values, and changes to the storage type updates the storage value. As a result, the corresponding utilization values are re-computed and the CPU Util, Memory Util, and Storage Util values are updated. For example, consider that the Lift-and-shift strategy is applied and the proposed AWS instance type is c4.8xlarge. The CPU Util shows 10%, Memory Util shows 50%, and Disk Util shows 12%. These values indicate that the target server is oversized and underutilized, and hence, you might want to select an instance type with a smaller configuration.
Operating System
Modify the default operating system that is running on the proposed server.
Storage Type and Storage
Specify the amount of storage that you need to buy for the proposed server.
To do so, select a storage type from the predefined list and specify the storage value.
Expand the following sections for specific points and steps per the public cloud provider:
- To save your changes, click Apply.
The total migration cost estimates for the corresponding cloud provider is updated.
- Ensure that the Simulation migration for <source server> window is open that displays the details of the proposed server that you want to modify.
4.Save a copy of the simulation results for offline use
Click Export and select Export to PDF or Export to Excel.
Cloud to cloud migration scenarios
Along with the scenario of migrating servers that are hosted in your on-premises infrastructure to the public cloud providers, the following migration scenarios are also supported:
Source cloud provider | Target cloud providers | Description |
---|---|---|
Hybrid (Mix of on-premises, AWS, Azure, Google Cloud) |
| For example, a business service might use some servers that are hosted on a public cloud while some servers that are still in the on-premises data center. |
AWS |
| You can choose to simulate the migration of a server between regions of the same cloud provider. For example, the instance type of your choice is unavailable in the current region, then you can migrate the server to a region where it is available. |
Azure |
| |
Google Cloud |
|
As shown in the following screenshot, all the servers that are selected for migration are hosted on AWS. The simulation results page shows the migration to the target cloud providers, including AWS.
The following screenshot shows the case of a hybrid scenario where some of the selected servers are hosted on a public cloud (AWS) while some servers are hosted in your on-premises infrastructure.