Leveraging Generative AI Assistants to Optimize Skytap TCO

Recently, I evaluated a cloud migration project in which the customer intended to transfer all their workloads from on-premises infrastructure to Kyndryl’s Skytap public cloud. The initial approach, a "like for like lift and shift," led to the Total Cost of Ownership (TCO) being lower for the on-premises setup compared to Skytap. This made me wonder if publicly available Generative AI tools using “web-based information” like Microsoft’s Copilot for Bing Chat AI Companion could help me to improve the Skytap solution.

My initial Copilot for Bing prompt was: "I am creating a solution to migrate on-premises AIX workloads to the Skytap public cloud using a 'lift and shift' strategy. What are some ideas for optimizing the total cost of ownership?" Copilot for Bing responded with the following:

“Migrating AIX workloads to the Skytap public cloud using a “lift and shift” strategy can be a complex process, but there are several ways to optimize the total cost of ownership (TCO). Here are some ideas:

1. Right-Sizing Resources: Ensure that the resources you allocate in the cloud match the actual needs of your workloads. Over-provisioning can lead to unnecessary costs. Regularly monitor and adjust resource allocation based on usage patterns.
2. Automated Scaling: Implement automated scaling to adjust resources dynamically based on demand. This helps in avoiding over-provisioning during low-demand periods and ensures availability during peak times.
3. Reserved Instances: If your workloads have predictable usage patterns, consider using reserved instances or long-term commitments. These can offer significant cost savings compared to on-demand pricing.
4. Cost Management Tools: Utilize cost management and monitoring tools provided by Skytap or third-party solutions.?These tools can help you track usage, identify cost-saving opportunities, and set up alerts for unusual spending patterns1.
5. Optimize Storage: Choose the right storage options based on your performance and cost requirements. Use tiered storage solutions to balance between high-performance and cost-effective storage options.
6. Network Optimization: Minimize data transfer costs by optimizing your network architecture. Use private connections where possible and reduce unnecessary data movement between on-premises and cloud environments.
7. Decommission Unused Resources: Regularly audit your cloud environment to identify and decommission unused or underutilized resources. This can help in reducing waste and lowering costs.
8. Leverage Skytap Features: Take advantage of Skytap’s built-in features for cost optimization, such as automated environment scheduling, which can shut down environments during non-business hours to save costs.
9. Training and Best Practices: Ensure your team is well-trained on cloud best practices and cost optimization strategies. This can help in making informed decisions and avoiding common pitfalls.
10. Hybrid Solutions: Consider hybrid solutions where some parts of the workload remain on-premises while others are moved to the cloud.?This can help in balancing costs and performance based on specific workload requirements2.

By implementing these strategies, you can optimize the total cost of ownership for your AIX workloads in the Skytap public cloud. If you have any specific concerns or need further details, feel free to ask!”

?

Copilot for Bing introduced several ideas, such as “right-sizing” resources, which are well-known FINOPS optimization strategies that I am familiar with. However, I was particularly interested in “8. Leverage Skytap Features”. So, my next prompt was, “Tell me more about Skytap's automated environment scheduling.” Copilot for Bing replied with:

“Skytap’s automated environment scheduling is a powerful feature designed to help you manage and optimize your cloud resources efficiently. Here are some key aspects:

1. Scheduling Actions: You can automate various actions such as starting, stopping, suspending, or saving environments.?This allows you to control when your environments are active, which can help reduce costs by ensuring resources are only used when needed1.
2. Recurring Schedules: You can set up recurring schedules to automate actions on a daily, weekly, or custom basis.?For example, you might start a test environment every weekday morning and suspend it at the end of the workday1.
3. Time Zone Adjustments: Schedules can be configured to run based on specific time zones, ensuring that actions occur at the right times regardless of where your team is located1.
4. Environment Management: You can create schedules that automatically save environments as templates, providing a backup or a starting point for new environments.?This is useful for maintaining consistent configurations and quickly deploying new instances1.
5. Auto-Shutdown for Inactive Environments: Skytap also offers options to automatically shut down or suspend environments that are inactive, helping to save resources and reduce costs2.
6. VM Sequencing: For complex environments, you can control the order in which virtual machines start up and shut down.?This ensures that dependencies are managed correctly, such as starting database servers before application servers2.

By leveraging these scheduling features, you can optimize resource usage, reduce operational costs, and ensure your environments are managed efficiently.?If you need more detailed guidance on setting up schedules, Skytap’s documentation provides comprehensive instructions12.”

?

In a matter of seconds, I discovered key aspects of Skytap’s automated environment scheduling and was provided with a link to Skytap’s documentation. ?At this point, I decided to step back. My next prompt was “How can I estimate costs before migrating?” Copilot for Bing responded with:

“Estimating costs before migrating your AIX workloads to the Skytap public cloud is crucial for planning and budgeting. Here are some steps and tools you can use to get an accurate estimate:

1. Use Cost Estimation Tools

• Skytap Cost Estimator: Skytap provides a cost estimation tool that helps you calculate the potential costs based on your specific workload requirements. This tool considers factors like compute, storage, and network usage.
• Azure TCO Calculator: If you’re using Skytap on Azure, you can use the Azure Total Cost of Ownership (TCO) calculator. This tool helps you compare the costs of running your workloads on-premises versus in the cloud1.

2. Assess Current On-Premises Costs

• Hardware and Infrastructure: Calculate the costs of your current hardware, including servers, storage, and networking equipment.
• Operational Costs: Include costs for power, cooling, physical space, and maintenance.
• Staffing: Consider the salaries and training costs for your IT staff.

3. Define Your Cloud Requirements

• Resource Needs: Determine the compute, storage, and network resources your workloads will require in the cloud.
• Usage Patterns: Analyze your workloads to understand their usage patterns. This will help you decide between on-demand and reserved instances.

4. Consider Migration Costs

• Initial Setup: Include costs for initial setup and configuration in the cloud.
• Data Transfer: Estimate the costs for transferring data from your on-premises environment to the cloud.

5. Factor in Additional Services

• Disaster Recovery and Backup: Include costs for disaster recovery and backup solutions.
• Security and Compliance: Consider any additional costs for meeting security and compliance requirements in the cloud.

6. Review Cost Management Tools

• Monitoring and Optimization: Use cost management tools to monitor and optimize your cloud spending. Skytap and Azure offer built-in tools for tracking usage and identifying cost-saving opportunities2.

Example Scenario

A detailed example scenario from Microsoft illustrates migrating AIX workloads to Skytap on Azure, highlighting the benefits of consumption-based usage plans and the ability to scale on demand34.

By following these steps and using the available tools, you can create a comprehensive cost estimate for migrating your AIX workloads to the Skytap public cloud. If you need more specific guidance or have any questions, feel free to ask!”

Copilot for Bing, my AI Companion, pinpointed specific tools for cost estimation in Skytap and Azure, along with recommending additional TCO factors. My next request was, “Can you provide a TCO template?” to which Copilot for Bing responded:

“Certainly! Here are some resources and steps to help you create a Total Cost of Ownership (TCO) comparison for migrating AIX workloads to Skytap:

1. TCO Template

Here’s a basic template to get you started:

TCO Comparison Template

Category

On-Premises Costs Skytap Costs Initial Costs

Hardware $ $

Software Licenses $ $

Migration Costs $ $

Operational Costs

Maintenance $ $

Power and Cooling $ $

IT Staff $ $

Scalability

Additional Hardware $ $

Disaster Recovery

Backup Solutions $ $

Software Licensing

License Management $ $

Performance & Reliability

Maintenance and Updates $ $

Security & Compliance

Security Measures $ $

Compliance Costs $ $

Long-Term Costs

Hardware Depreciation $ $

Future Upgrades $ $

Total Costs $ $

2.????? ?Steps to Use the Template

1. Identify Costs: Fill in the costs for each category based on your current on-premises setup and the estimated costs for Skytap.
2. Compare: Analyze the differences in costs to understand potential savings.
3. Adjust: Modify the template to include any additional costs specific to your organization.

By using these tools and templates, you can create a comprehensive TCO comparison to help you make informed decisions about migrating your AIX workloads to Skytap. If you need further assistance or specific details, feel free to ask!”

The template was practical and could be seamlessly integrated into a spreadsheet. Additionally, I was impressed that Copilot for Bing offered “steps to use the template”. My last test aimed to establish a TCO "rule of thumb". I asked, “What is the percentage TCO of Skytap compute, storage, and network costs?” Copilot for Bing replied with:

“The percentage of Total Cost of Ownership (TCO) attributed to compute, storage, and network costs can vary based on the specific workload and usage patterns. However, for a typical “lift and shift” migration strategy, the breakdown is often as follows:

1. Compute Costs: This usually represents the largest portion of TCO, typically ranging from 50% to 70%. Compute costs include the virtual machines (VMs) or instances required to run your workloads1.
2. Storage Costs: Storage costs generally account for 20% to 30% of the TCO. This includes the cost of storing data in the cloud, such as block storage, object storage, and any associated backup storage2.
3. Network Costs: Network costs are usually the smallest portion, ranging from 5% to 15%. These costs include data transfer charges, VPN or direct connection fees, and any additional network services like load balancing3.

These percentages can vary based on factors such as the specific cloud provider’s pricing model, the nature of your workloads, and your usage patterns. To get a more accurate estimate for your scenario, consider using Skytap’s cost management tools and cloud provider calculators to model your expected usage and associated costs.

If you have any more questions or need further details, feel free to ask!”

Referring to the hyperlinks in Copilot’s response, I was surprised to see that Copilot for Bing's response integrated sources from both Microsoft and NetApp. This illustrates a situation where Generative AI may be embellishing the truth when producing a response.

?As my AI Companion, I was impressed by how efficiently Copilot for Bing provided insights to optimize the Skytap solution. However, as a knowledge worker, I remain mindful when incorporating generative AI in my tasks, recognizing that it does not replace my creativity. I utilize the Generative AI tools approved by Kyndryl, my employer, and ensure that I do not share any sensitive or confidential information while using them. I rely on my expertise and judgement to interpret and use AI-generated content appropriately. I cross-check crucial facts and data to confirm their accuracy.

?I'll keep using Microsoft's Copilot for Bing Chat AI Companion whenever it seems helpful. Please reply to this post—I’m eager to hear your thoughts on using Generative AI for solution development.