Powerleader Helps New Energy Technology Build A Hyper-converged Information Platform
Applications
After More Than 20 Years Of Development, Blade Has Continuously And Deeply Grasped Market Demand And Accumulated A Large Number Of Customer Resources And Rich Service Experience.
Case Background
ATL, Headquartered In Hong Kong, Is A High-tech Enterprise Dedicated To The R&D, Production And Marketing Of Rechargeable Lithium-ion Battery Cells, Packaging And System Integration. Among The Global Professional Lithium Battery Manufacturers, Its Technology, Production Capacity And Sales Are In A Leading Position. The Lithium Batteries Produced By ATL Are Widely Used In Various Consumer Electronic Products Such As Laptops, Mobile Phones, Digital Players, Digital Cameras, Portable Video Recorders, Power Tools, Etc. At The Same Time, ATL Is Actively Expanding The Electric Vehicle And Energy Storage System Markets.
Case Analysis
The Challenges Faced By The Existing Data Center Of New Energy Technology Include:
A Large Number Of Servers, Which Makes Management Increasingly Complex;
A Long Launch Cycle And Slow Deployment Of New Business Systems;
Poor Scalability Of SAN/NAS Storage, Which Cannot Support The Performance Requirements Of New Businesses;
New Businesses Are Moving Towards The Internet, And Traditional Architectures Cannot Achieve Linear Scalability;
Application Systems Lack High Availability Protection; And
Data Center Space Resources Are Limited.
The Construction Of New Data Centers Includes The Implementation Of Many Information Applications. Correspondingly, Server Rooms And Storage Devices Will Also Be Used In Large Quantities, And With The Expansion Of Applications And Services In The Later Stages, The Investment In Servers And Storage Devices Will Inevitably Become Larger And Larger. On The One Hand, The Pressure And Cost Of Managing Hardware Infrastructure Will Continue To Increase; On The Other Hand, Due To The Diversity Of Applications, Servers And Storage Are Difficult To Effectively Integrate, And The Resource Usage Of Servers Is Far Lower Than Their Actual Processing Capabilities, Making It Difficult To Fully Utilize Computing Power And Storage Capacity.
Solution
Implementing Virtualization/cloud Computing Data Centers Can Effectively Integrate Server And Storage Resources To Form A Computing Resource Pool, Dynamically Allocate Computing Resources According To The Actual Needs Of Various Applications In The New Generation Of Data Centers, Maximize The Use Of Existing Servers And Storage Devices, And Effectively Manage And Monitor Data Center Hardware Equipment.
1. Architecture Design
The Hyper-converged Architecture Plays The Role Of Computing Resource Pool And Distributed Storage Resource Pool In The Data Center, Greatly Simplifying The Infrastructure Of The Data Center, And Through Software-defined Computing Resource Virtualization And Distributed Storage Architecture, It Achieves No Single Point Of Failure, No Single Point Of Bottleneck, Elastic Expansion, And Linear Performance Growth. At The Virtualization Layer, You Can Freely Choose The Brand Of Hypervisor, Including Vmwarevsphere, Microsofthyper-v And KVM; And Through A Simple And Convenient Management Interface, You Can Achieve Unified Monitoring, Management And Operation Of Computing, Storage, Virtualization And Other Resources At The Data Center Infrastructure Layer.
- Computing Resource Design
The Modular Data Center Based On The Powerleader Hyper-convergence Architecture Is Composed Of Multiple Converged Nodes. Figure 2.2 Below Shows A Standard Hyper-convergence Device That Only Occupies 2 Rack Units (2u High). Each Standard Device Contains Four Independent Converged Nodes (Node), And Each Node Is An Independent X86 Server. Each Device Can Provide 4 Standard 2-way Intel CPU X86 Servers And A Maximum Storage Capacity Of 32tb. The Computing Resource Pool Of The Powerleader Hyper-convergence System Is Realized Through X86 Server Virtualization, Which Can Support Hypervisors Such As Vmware Vsphere, Microsoft Hyper-v And KVM, As Shown In Figure 2.3. The Computing Resource Pool Is Formed At The Virtualization Hypervisor Layer To Provide Different Service Qualities And Capabilities For The Virtual Machines Of The Business System, Including Virtualization Features Such As High Availability (High Availability), Fault Tolerance (Fault Tolerant), Online Migration (Live Migration/vMotion), And Dynamic Resource Load Balancing (Distributed Resource Scheduler). - Storage Resource Design
The Distributed File System (Celph) Provided By The Powerleader Hyper-converged System Can Form A Group Of Nodes In A Cluster Into A Unified Distributed Storage Platform. Celph Is A Centralized Shared Storage For X86 Virtualization Platform Software. It Works Like Any Other Centralized Storage Array And Provides Simpler And More Convenient Storage Management. There Is No Need To Configure LUN, Volume, Or Raid Group Like Traditional Centralized Storage. The Distributed Storage Architecture Of The Powerleader Hyper-converged System Not Only Provides The Same Traditional Storage Capabilities, But Also Provides More Capabilities. It Provides Snapshots, Cloning And Other Mechanisms For Virtualization, And The Data Layer Implements Local Priority Access, Storage Tiering And Other Performance Mechanisms. It Compresses And Deduplicates Data To Increase Storage Available Capacity, Provides Storage Reliability With More Than Two Redundant Data, And Automatically Increases Or Decreases Node Data Distribution. It Has Data Self-recovery Capabilities When A Node Goes Down Or A Disk Is Damaged.
For Network Isolation, Network Virtualization VXLAN Technology Can Also Be Used. VXLAN Network Protocol Is An Extended Version Of VLAN Protocol. VXLAN Network Can Cross Physical Boundaries, Thereby Optimizing Computing Resource Utilization Across Discontinuous Data Centers And Clusters. Vxlan Separates Logical Network And Physical Topology From Each Other And Uses Ip Technology, So The VXLAN Network Can Be Expanded Without Reconfiguring The Underlying Physical Network Equipment. Because Of This, There Is No Need To Spend A Lot Of Time Planning How To Allocate VLANs And Manage The Rapid Increase In The Number Of VLANs. There Are Multiple Network Requirements On Each Physical Node, Including Internal Communication Network, Management Network, Production Network, Etc. Therefore, Each Node Needs To Be Configured With Multiple Network Cards. The Network Design Is As Follows:
type | design | Remark |
Internal communication network | 10Gb Ethernet Dual link redundancy | Each node is connected to two 10G switches through two 10G links to ensure the redundancy of network devices and links. It is recommended that users use 10G network to interconnect physical nodes. When intensive write IO occurs , the 10G network can ensure sufficient The bandwidth satisfies the IO synchronization traffic between nodes. |
Communication network between client and server virtual machines, virtualized servers External service network | 1Gb/10Gb Ethernet, dual link redundancy | Each node is connected to two 1000/1000 Gigabit switches through two 1000/1000 Gigabit links to ensure the redundancy of network devices and links . When users access a virtual server to provide services externally, Thousand/10 Gigabit links can be isolated from back-end storage traffic. |
Hardware Management Network (IPMI) | 1Gb Ethernet | Each node has an independent Gigabit link to connect to a dedicated management network, achieving Storage network separation can maximize management flexibility. |
- Backup and disaster recovery design
The PowerLeader hyper-converged platform comes with a storage layer and a virtual machine-based backup and recovery function called TimeStream. Users can set different backup policies for each virtual machine, including backup plans and backup retention periods. TimeStream will automatically back up the virtual machine through storage snapshots. All TimeStream snapshots are based on the storage layer. Unlike snapshots at the virtualization layer (such as VMware vSphere), snapshots at the storage layer do not affect the performance of the virtual machine and are completely transparent to the virtualization software.
Customer Benefits
With The Boya Virtualization Infrastructure, While Ensuring High-speed Access And High Reliability Of User Data, The Traditional Centralized Storage Architecture Is No Longer Needed, Avoiding Performance Problems That Were Overlooked In The Early Design Stage During Future Operation. According To The Current Status Of Server And Storage Usage, It Is Recommended To Use Hyper-convergence To Support Existing Applications. In The Future, As The Business Develops, It Can Be Easily Expanded According To Nodes To Avoid The Problem Of Excessive One-time Investment In The Initial Construction, And Then Discovering Performance Problems Later And Requiring Additional Investment. Boya Hyper-convergence System Combines The Advantages Of Converged Infrastructure, Horizontal Expansion Architecture And Software-defined Storage To Provide An Excellent Data Center Virtualization Experience, And Its Cost Is Only A Fraction Of Traditional Servers And Storage. N Hyper-convergence Full Cluster Is A Computing And Storage Infrastructure That Can Be Scaled Out, Which Enables Organizations To Achieve Data Center Virtualization Without A Storage Network (SAN Or NAS). Blade Hyper-convergence Is Designed Specifically For Virtualization And Can Provide Comprehensive Computing And Storage Capabilities, As Well As Enterprise-level Performance, Scalability, Availability And Data Management Functions. It Uses Industry-standard Hardware Components Such As Intel Chips And Solid-state Drives, As Well As Market-leading Advanced Management Programs To Provide Out-of-the-box Solutions, Making Data Center Virtualization Extremely Simple And Effective.
