Showcases
.svg){kind=logo}
Author／Tech Service IV-III Solution Architect Champ Huang,
Tech Service IV-III Manager Roy Syu,
Tech Service IV Manager Alvin Su.
In this article, we will introduce how MetaAge Digital AWS MSP team helps OnIT to build deep learning solutions for customers on AWS. MetaAge digital agency AWS cloud service brand, provides technical services and 24/7 MSP hosting services, and further obtains the AWS MSP partner qualification. In addition to the MSP maintenance team that provides a 24/7 MSP maintenance team, MetaAge Digital's AWS technical team also includes technical engineers with rich experience in traditional servers, networks and databases, and has AWS software development consultants, so whether customers want to deploy AWS services, or further value-added services such as managed maintenance, API integration, cloud integration, etc. MetaAge Digital AWS team has complete professional capabilities, and has Relevant MetaAge Digital agency brands have customized development departments which can work together to customers and provide customers with one-stop multi-faceted complete services.
MetaAge Digital's AWS technical team assisted OnIT Technology to build deep learning solutions for customers on AWS, and hosted its cloud environment by MetaAge Digital to optimize customers' experience at multiple levels, including:
1. Computing Solutions
2. System optimization
3. Maintenance Agility
4. Information Security
5. Guaranteed availability
6. Cost savings
MetaAge Digital's AWS MSP maintenance team also assisted OnIT to introduce the next generation monitoring and work order system, providing customers with fully managed and fully monitored MSP integration services.
About OnIT Technology：OnIT Technology is a comprehensive system integration service provider, providing diversified IT solutions for Taiwanese enterprises. Its customer industries mainly cover high-tech, semiconductor, medical, traditional manufacturing, telecommunications, and financial industries, and cooperate with many major global leaders. Data center solution manufacturers have established partnerships, and the technology includes various professional information fields, providing excellent professional technology and the best service quality.
In recent years, with the Covid-19 epidemic and the wave of virtual currency, there have been chip shortages all over the world. Many companies have been struggling to purchase corresponding machines, and the computing equipment required for deep learning is one of the biggest victims. OnIT Technology (hereinafter referred to as OnIT) has very rich experience in building a complete deep learning system platform for customers and has an excellent technical team to maintain its operation, but the delay in delivery and costs caused by the shortage of chips. In the rising situation, a lot of troubles had been encountered when they try to construction and purchase their customers' equipment of local host required for deep learning. At this time, MetaAge Digital (hereinafter referred to as MetaAge) and OnIT have been in partnership many years., after many discussions with the MetaAge AWS consultant team, they decided to adopt the cloud solution to provide solutions for customers, and hope to improve the company's business agility and maintenance efficiency through the additional service value of AWS.
(Image provide by AWS official website: https://aws.amazon.com/tw/machine-learning/infrastructure/ © Amazon Web Service)
(Service Used: Amazon EC2 G4 Instance Type) Through the AWS EC2 solution, you can immediately have deep learning computing services with better value. The G4 Instance Type selected by MetaAge to help OnIT is the most economical GPU Instance in the cloud industry, and can provide up to 1.8 TB of NVMe (Non-Volatile Memory Express) SSD. store. Taking the Instance Type of g4dn.xlarge that was decided to be deployed, the GPU of this Instance is a new generation of NVIDIA T4 GPU with 16 GB GPU Memory and 1 x 125 NVMe SSD. OnIT can use this specification to promote the computing service originally built for customers on the local side, without changing too many settings for this, it can reach the original service level, and can immediately provide customers with corresponding service content.
OnIT's users have the requirement of remote desktop connection to Linux system, so in addition to cooperating with OnIT, MetaAge established Ubuntu 20.04 AMI, the operating system required for deep learning, and customized it for RDP connection. We use the Xfce desktop and use the LightDM lightweight fast display management platform to improve the operation speed let the operation of the desktop is smoother.
(Service Used: Elastic File System, EC2 User data, System Manager) The customer has the demand for NFS File Storage, and because the related system requires high-performance computing, the IOPS of the hard disk needs to reach 10,000, and the capacity needs to be 20TB. This project uses Amazon Elastic File System after evaluation by the two teams of OnIT and MetaAge to meet the relevant requirements, achieve the service level traditionally provided, and improve the agility of application. The customer needs to use more than 10 Instances, so MetaAge MSP team and OnIT technical team help customers create a pre-written User Data script during maintenance and operation, so that OnIT engineers help customers deploy Instances quicker than usual. In addition, in the management of Instances, we manage the established Instances through the System Manager Fleet Manager, so that the engineers can quickly control all the established machines and conduct maintenance.
(Service Used：CloudFormation) For the part of Ubuntu RDP connection, OnIT adopts the remote secure connection product of MetaAge digital agent, and automatically deploys it on EC2 through CloudFormation to provide services for customers. The connection method is through Outbound Https encrypted connection and Clients communicate with each other, and users can directly connect to Instance resources through the Internet without connecting to the public IP, and the security of the connection is guaranteed.
(Service Used: AWS Capacity Reservations、Instance Saving Plans; MetaAge Next Generation Monitoring, Billing and Ticket System) The maximum G Instance Type selected by the customer will reach 120vCPUs. In the case of an instance that needs to be restarted, if it encounters a peak computing period, it is easy to fail due to insufficient capacity. Therefore, MetaAge helped OnIT to use AWS Capacity Reservations establishes a capacity reservation mechanism to ensure that Instances are available at any time to the highest degree to maintain service availability. In addition, because a small number of Instances of customers will need to be powered on for 24 hours and continue to be used for one year, we evaluate these machines through MetaAge's next-generation accounting system and recommend OnIT to use Instance Savings Plans to promise computing usage, which greatly reduce computing costs. In the end, MetaAge MSP introduced the next generation monitoring and work order system for OnIT Technology. According to customer needs and the occurrence of monitoring events, relevant SLAs are immediately intervened by MSP personnel to ensure the availability and continuity of deep learning services and complements the solid technical team of OnIT to provide more convenient and brand-new deep learning services.
Through MetaAge MSP team and OnIT provide customers with:
1. Availability
Through MetaAge MSP and imported monitoring and work order system, events that affect availability can be detected early, reducing MTTD (Mean Time to Detect), MTTR (Mean Time to Repair) and improving availability.
2. Consistency
There are scripted and proven operations when creating an instance, which reduces the possibility of human error compared to manually entering instructions.
3. Cost savings
Purchasing the Instance Saving Plan enables OnIT to save up to 22% of the cost on Instances that need to run continuously for 24 hours.
MetaAge has rich experience in server, network, database management, and has a deep technical architect team and maintenance team. It is the best consultant and service team for enterprise cloud services. Whether customers want to deploy AWS services, Cloud Managed Service, Infrastructure-as-code, API integration, cloud integration and other value-added services, MetaAge Digital has complete professional capabilities, All can meet multi-faceted needs in one stop. MetaAge Digital's AWS MSP team introduced the next-generation monitoring system to provide customers with fully managed and fully monitored MSP integration services. MetaAge officially obtained the status of AWS MSP (Managed Service Provider) in April of 2022. For AWS, we provide cloud consulting, cloud hosting and other services. If you have any questions or needs on AWS, please feel free to contact us!
Contact Information：
Phone: 080-000-8669
Email: aws@metaage.com.tw/
Line@: @metaage_msp
● MetaAge AWS MSP Next Generation Monitoring System Introduction
● Amazon EC2 G4 Instance Industry's most affordable GPU Instance for machine learning inference and graphics-intensive applications
