Statement of Work (SOW)
Service Overview
As your cloud storage needs grow, you may encounter performance bottlenecks. For instance, you might require greater storage elasticity during service peaks or faster AI model training as data and parameters increase. Industries operating across the Internet have even more demanding storage requirements, including performance, security, and usability. Nevertheless, configuring storage resources to optimize service support can be challenging.
To satisfy your storage needs in various situations, Huawei Cloud has introduced the Cloud Storage Optimization Service (CSOS). This service evaluates your storage performance and offers optimization solutions to help you make the most of cloud storage while minimizing expenses.
Service Content
- Cloud Storage Performance Evaluation and Optimization for Large AI Model Training
Specifications
Content
Applicable Scenario
SFS Turbo Performance Evaluation and Optimization for Large AI Model Training - Basic Package
The performance of your cloud storage for large AI model training will be evaluated, and optimization solutions will be provided. These include selecting the SFS Turbo file system types, designing a three-level cache solution, and generating an AI Turbo SDK. Additionally, data interworking solutions will also be provided to speed up the training process.
If you face performance bottlenecks while using cloud storage for AI model training and seek faster and more cost-effective solutions, try CSOS.
SFS Turbo Performance Evaluation and Optimization for Large AI Model Training - Incremental Package
- Cloud Storage Performance Evaluation and Optimization
Specifications
Content
Applicable Scenario
EVS Performance Evaluation and Optimization - Basic Package
The performance of cloud storage (EVS) will be evaluated, and optimization solutions will be provided to improve reliability in different scenarios. These solutions include configuring image cache, provisioning elastic resources in batches, removing local disks, and designing high-performance storage solutions for big data and databases. Solutions for selecting EVS disk types will also be provided to meet diversified requirements.
We offer performance evaluation and optimization solutions to address performance issues with cloud storage (EVS). For example, if you come from a medium- or large-sized enterprise, financial institution, or a government agency that suffers from a high local disk failure rate, we offer reliable, high-performance cloud storage solutions. If you work in industries such as video rendering and social media, we can provide you with robust cloud storage to handle sudden bursts of traffic, batch provision resources, and to scale out storage.
EVS Performance Evaluation and Optimization - Incremental Package
Prerequisites
- Customers have migrated workloads to Huawei Clouds and are using Huawei Cloud storage products.
- Customers are highly cooperative. The owner on the customer side takes the lead and assigns personnel to conduct surveys and analysis, and formulate plans.
- Before providing customers with the cloud storage performance evaluation reports, Huawei must obtain customers' authorization to view their behavior data.
- Both parties agree on the service objectives and customers pay for the service.
Service Scope
- Applicable Scope
During the entire process, Huawei Cloud provides the following services:
- SFS Turbo performance evaluation and optimization for large AI model training
The performance of your cloud storage for large AI model training will be evaluated, and optimization solutions will be provided. These include selecting the SFS Turbo file system types, designing a three-level cache solution, and generating an AI Turbo SDK. Additionally, data interworking solutions will also be provided to speed up the training process.
- EVS performance evaluation and optimization
The performance of cloud storage (EVS) will be evaluated, and optimization solutions will be provided to improve reliability in different scenarios. These solutions include configuring image cache, provisioning elastic resources in batches, removing local disks, and designing high-performance storage solutions for big data and databases. Solutions for selecting EVS disk types will also be provided to meet diversified requirements.
- SFS Turbo performance evaluation and optimization for large AI model training
- Inapplicable Scope
Services implemented by customers, including but not limited to:
Integrating different systems
- Regions
Service Process
|
Phase |
Description |
|---|---|
|
Requirement survey |
Clarifying service objectives: Huawei Cloud conducts surveys on the customer's storage service statuses and basic requirements, interviews the customer, and agrees with the customer on service objectives. Formulating solutions: Huawei Cloud aligns detailed solutions with the customer, including the service content, process, schedule, and expense, to ensure that the solutions meet the customer's requirements and are feasible. |
|
Evaluation and diagnosis |
Huawei tests the performance of the cloud storage system that is being used by the customer, analyzes the performance bottlenecks, and outputs a diagnosis report. |
|
Solution design |
Huawei clarifies existing problems and provides corresponding solutions through in-depth communication with the customer. |
|
Service acceptance |
Customers accept the solutions. Huawei Cloud hands over related documents. |
Service Deliverables
- Cloud Storage Performance Evaluation and Optimization for Large AI Model Training
Specifications
Deliverable
Acceptance Report
SFS Turbo Performance Evaluation and Optimization for Large AI Model Training - Basic Package
Huawei Cloud Large AI Model Storage Solution Planning and Design
Cloud Storage Optimization Service Acceptance Report for the XX Project
SFS Turbo Performance Evaluation and Optimization for Large AI Model Training - Incremental Package
- Cloud Storage Performance Evaluation and Optimization
Specifications
Deliverable
Acceptance Report
EVS Performance Evaluation and Optimization - Basic Package
Huawei Cloud EVS Storage Solution Planning and Design
Cloud Storage Optimization Service Acceptance Report for the XX Project
EVS Performance Evaluation and Optimization - Incremental Package
Responsibility Matrix
- Huawei Cloud and Customers
- Negotiate and confirm specific requirements and objectives.
- Negotiate and confirm project management plans.
- Negotiate, confirm, and review solutions.
- Sign contracts.
- Huawei Cloud
- Receives customer requirements and designates experts to help implement and guarantee the service delivery onsite or remotely.
- Develops a service plan and quotation based on the service items for customers to review and confirm.
- Implements the service as planned and compiles deliverables.
- Provides the list of deliverables based on the selected service items after services are all complete.
- Specifies a project owner and notifies customers of any personnel changes three working days in advance.
- Uses the authorized data only for this service and not for any other purposes.
- Customer
- Assigns a project owner to assist Huawei Cloud in implementing services. The project owner is responsible for arranging and coordinating personnel, resources, or other matters between the two parties, and reviews and accepts the services provided by Huawei Cloud.
- Provides service system-related information, including but not limited to application architecture, deployment architecture, resource quantity, and performance.
- Responsibility Details
This table provides responsibility matrix examples and can be modified as needed.
R: responsible; S: supportive.
- Cloud Storage Performance Evaluation and Optimization for Large AI Model Training
No.
Process
Content
Huawei
Customer
1
Requirement survey
Specify service objectives.
R
R
2
Formulate solutions.
R
S
3
Performance evaluation
Evaluate performance.
R
S
4
Solution design
Formulate performance optimization solutions.
R
S
5
Communicate with the other party and confirm the solutions.
R
R
6
Service acceptance
Accept the solutions and related documents provided by Huawei Cloud.
S
R
7
Guide the customer team to implement the performance optimization solutions.
R
S
- Cloud Storage Performance Evaluation and Optimization
No.
Process
Content
Huawei
Customer
1
Requirement survey
Specify service objectives.
R
R
2
Formulate solutions.
R
S
3
Performance evaluation
Evaluate performance.
R
S
4
Solution design
Formulate performance optimization solutions.
R
S
5
Communicate with the other party and confirm the solutions.
R
R
6
Service acceptance
Accept the solutions and related documents provided by Huawei Cloud.
S
R
7
Guide the customer team to implement the performance optimization solutions.
R
S
- Cloud Storage Performance Evaluation and Optimization for Large AI Model Training
Acceptance Criteria
- Cloud Storage Performance Evaluation and Optimization for Large AI Model Training
For the SFS Turbo performance evaluation and optimization packages for large AI model training, Huawei Cloud submits standard deliverables based on service sub-items. Customers confirm the acceptance on the Huawei Cloud official website console, or signs and seals the Cloud Storage Optimization Service Acceptance Report for the XX Project onsite as the basis for service acceptance.
- Cloud Storage Performance Evaluation and Optimization
For the EVS performance evaluation and optimization packages, Huawei Cloud submits standard deliverables based on service sub-items. Customers confirm the acceptance on the Huawei Cloud official website console, or signs and seals the Cloud Storage Optimization Service Acceptance Report for the XX Project onsite as the basis for service acceptance.
Feedback
Was this page helpful?
Provide feedbackThank you very much for your feedback. We will continue working to improve the documentation.See the reply and handling status in My Cloud VOC.
For any further questions, feel free to contact us through the chatbot.
Chatbot
