catalogue
First, overview
II. Functional requirements
(A), Modularization
(2) Integration of hardware and software
(3) Automatic scaling
(Iv) Failover
(5) Load balancing
(Vi) Complete cloud lifecycle management and virtual machine lifecycle management.
(7) Easy to expand and easy to upgrade
(8) Easy distribution, easy operation and maintenance
(9.) High safety and high reliability
3. System design
(I) Architecture design
(2 i) Hyper-fusion resource design
(3) Design of the storage resource pool
4. Main equipment parameters
The planning and construction of the data center for the safety of production business and enterprise business services. The data center is composed of servers, storage equipment, backup all-in-one machine, switch and other equipment.
The cloud computing platform assigned by each node is a complete computing module, integrating the hardware data and software resources such as network, storage, server, data acquisition and transmission, and power distribution units required for the application.
(2) Integration of hardware and software
The storage, network and server resources are reasonably integrated and adjusted, and make full use of hardware resources and software resources to form a flexible cloud environment, and provide a high-density virtual server environment.
The cloud platform can automatically size the required resources according to the application configuration. When the application's virtual machine load is excessive due to insufficient initial resource allocation, the system can automatically expand the same configured virtual machine to realize load sharing, and can carry out resource recovery when the load drops.
Provides a comprehensive set of high-availability features, including virtual machine and physical machine failover capabilities, to ensure that the mission-critical application software is functioning properly.
The cloud platform dynamically allocates service resources according to the needs of the application, realizes the collaborative work and parallel processing of multiple servers, greatly improves the server performance and makes full use of network resources.
(Vi) Complete cloud lifecycle management and virtual machine lifecycle management.
Cloud platform provides from application planning, installation, deployment, configuration, monitoring, change and so on a set of full production cycle of cloud management portal and for each virtual machine to provide complete virtual machine life cycle management, provide virtual machine creation, start, pause, restore, sleep, restart, shutdown, close off the power, modify, delete, query, resource recovery, and other functions.
(7) Easy to expand and easy to upgrade
Since the user only knows the interface of the application service and does not know the implementation process, this is equivalent to adding an intermediate layer to the private cloud and the user before. Changing the back end of the private cloud will not pass the impact to the front end and will not affect the use of users. The expansion, maintenance and upgrade of private cloud space brings flexibility, making the back end changes least affected.
(8) Easy distribution, easy operation and maintenance
Cloud platform using virtualization technology to break the traditional system of "one server an application", improve the utilization of software and hardware, realize the high availability of hardware and software and hardware resources, easy to manage, easy to maintain, can let the data center dynamically improve the performance and efficiency of IT infrastructure, realize the application of rapid deployment, virtual machine, backup rapid recovery, application upgrade test, quick back after upgrade failure, centralized performance monitoring and alarm function, maintain the continuity of the business. At the same time, the operation and maintenance personnel can conduct batch operation on the virtual machine in the cloud platform through the batch modification function, and the operation and maintenance personnel can regularly obtain the operation and maintenance log files in a certain period of time from the platform software with one click to facilitate archiving and reporting.
(9.) High safety and high reliability
The cloud platform is built for a single customer to use alone, thus enabling the most effective control over data security. The private cloud platform for the coal industry is deployed in a data center connected to a coal private network, so customers' internal employees reflect high availability when accessing private cloud applications.
The data center architecture is shown in Figure Figure 5 – 15.
Figure 5-15 Data center architecture diagram
Two core switches are deployed in the data center, connecting the industrial network and the management network core switch to the server, and the server access switch provides gigabit optical ports, connecting to the server service network through the optical port and the server management network through the optical port. The server area provides a hyperconverged server to provide computing and storage resources for cloud computing; all servers connect the switch as the service network through the Gigabit switch as the management network; through 2 * 10G, the interface connects the switch as the storage network to pool the hard disk of the hyperconverged server to provide storage resources. Deploy a single all-in-one backup machine at the back end to provide disaster recovery resources.
(2 i) Hyper-fusion resource design
The data center virtualization resource pool construction is mainly given priority to with super fusion architecture, according to the business scale choose server deployment business system, build on the basis of super fusion architecture of a new generation of super fusion resources cluster, to, in line with the modern application development framework in high performance, high reliability, flexible extension and expansion, simplify infrastructure, management requirements, and through standardization, realize the future in the automation failover, capacity, disaster, comprehensive optimization data center software and hardware assets, integrated intelligent operations of information goals, so as to realize the rapid response to the business.
As a computing and storage resource platform of the cloud platform, the hyper-fusion system provides resource support for the production, office of the entire coal mine and various business systems of the smart park. The whole hyper-fusion system adopts 10 computer-based storage fusion and node servers.
The hyperfusion node adopts the standard 2U rack server. Each server is equipped with 2.2GHz / 24-core CPU, 256G memory and 48TB bare capacity.10 units can provide 480 physical CPU cores, super fusion bottom layer consumption 48 cores, 432 cores, can provide 864 vCPU; 2560G memory, 480TB bare capacity, bottom layer consumption 18T, according to EC reliability deployment calculation, can provide 248TB available capacity. The ultra-converged resource pool can provide 864 vCPU, 900G memory, 248TB storage capacity, 8 vCPU, 16G memory, 500G capacity, and about 110 virtual machines, carrying the production network, management network and various business systems in the park.
(3) Design of the storage resource pool
This data center adopts distributed shared storage solution, which is mainly used for performance instance database resource pool, performance virtualization computing resource pool and backup storage requirements.
The data center hyper-fusion system uses server hard disk resources to provide distributed storage, 10 hyper-fusion servers deploy 4TB, 80 data disks, 10 NVMe cache acceleration disks above 3.84TB, can provide a total of 480TB of bare storage capacity, reliability deployment according to EC, a total of 248TB of available capacity.
Configure the backup system to back up the business system and stored data. The backup adopts LAN-BASE backup, connects with the storage and hyper-fusion system through the gigabit network to realize data backup, and realizes the non-agent backup of the hyper-fusion system and database or other system equipment. Backup equipment covers backup authorization and backup media hardware.
| order number | device name | Technical parameter requirements |
| 1 | Super-fusion all-in-one machine | 1. Through the construction of X 86 server node, computing storage fusion is realized within the same node, without the need for external SAN storage, the storage system is distributed Server SAN architecture, can be configured with 2 or 3 copies to meet different reliability requirements of business scenarios; 2. Support the storage nodes to install virtualization software, which can provide both virtual machine service and storage business; 3. Support horizontal expansion. When more computing and storage resources are needed, it only need to expand the server, that is, to realize the synchronous expansion of computing and storage resources; 4. Support hardware automatic discovery and automatic configuration, without manual participation; 5. Support mainstream database deployment in the industry, including but not limited to Oracle, Gbase, Renmin Cang, Dameng, PolarDB, etc.; 6. Support the monitoring and management of computing, storage, switches, virtualization platform, etc. in the unified management interface; 7. Support one-click or regular automatic output of system health inspection report on the unified graphical interface, including CPU, memory, HDD, SSD, RAID card and other hardware status, the health status of virtualization platform, storage software, management software and other components, to facilitate the active identification of potential risks; 8. Support the one-click log collection function on the unified graphical interface, and can quickly collect all the required log information, including hardware, virtualization platform, storage software and management software; 9. The computing nodes of a single cluster (HA resource pool) can be expanded to 128 units; 10. Support virtual machine resource adjustment, according to the actual needs to modify the attributes of the virtual machine, including the number of vCPU, memory size, the number of hard disks and the number of network cards; 11. Support the CPU, memory and storage of virtual machines to meet the performance requirements of different applications; 12. Support memory bubbles, memory exchange, memory sharing and other functions, to realize memory reuse allocation, improve resource utilization; 13. Virtual switch-level user state exchange technology (OVS + DPDK), which support high-performance network forwarding, improve data processing performance and throughput, and improve the work efficiency of data plane applications; 14. Distributed storage software is built on x 86 / ARM standard hardware, non-open source software development, such as open source Lustre and Ceph software cannot be used, and high scalability and redundancy to achieve decentralized architecture and data redundancy technology; 15. Under the full SSD, configuration and SSD + HDD hybrid configuration, EC (Erasure Cod e) algorithm is supported to realize the redundant data storage, supporting 2 + 2,4 + 2,6 + 2,8 + 2 multiple redundant configuration; 16. Support EC shrinkage. When the node fails, automatically adjust the EC ratio to ensure that the data reliability is not degraded; |
| 17. Allow 2 nodes without data loss and the storage utilization rate up to 80% 18. Support global adaptive redeletion and compression, automatically switch between online and later redeletion according to the business load; 19. Support multiple storage resource pools in a single storage cluster; support graphical interface to divide storage resource pool, and each storage resource pool is a fault domain to ensure reliability; 20. When the disk or storage node fails, the system can automatically reconstruct the data. Without manual intervention, the data reconstruction speed should be <15 minutes per TB; provide corresponding proof data; 21. Support volume snapshot and roll-back, the maximum number of snapshots supported by a single volume is not less than 2048, and the impact of snapshot on host business performance cannot exceed 5%; the snapshot should be based on ROW mode and support second-level snapshot; 22. Support disk sub-health management function: support regular detection of disk SMART, information, judge the disk sub-health situation (hard disk sector remapping number exceeds the threshold, read error rate statistics exceeds the standard, slow disk), and isolate and alarm before disk damage; 23, Support for SSD wear life identification, Early warning and isolation processing; 24, the required hyperfusion software and 20 CPU authorization; 25, hyperfusion all-in machine hardware configuration is as follows: CPU: 2, Main frequency is 2.2GHz, 24 cores per processor; memory: 24 memory slot bits, Configuring an 8 * 32G set of memory, Single requirement of 32G; Hard disk configuration: 2 blocks of 600G SAS 10K HDD, 1 block, 3,200 G NV M E, 8 pieces of 4T mechanical disks; Network interface configuration: 410GE optical ports (including multi-mode module). | ||
| 2 | core switch | 1. Forwarding performance: exchange capacity 4.8 Tbps, package forwarding rate 1600 Mpps; 2. Hardware specifications: 1U in height, fixed interface switch, 6100GE optical ports, and 48 10GE optical ports; 3, the actual configuration requirements: dual power supply; configuration of 21 1 trillion multi-mode optical modules; 4. Second 2 functions: support Access, Trunk and Hybrid; support QinQ; support dynamic MAC, static MAC, and black hole MAC items; 5. Three-layer functions: support IPv 4 dynamic routing protocols such as RIP, OSPF, ISIS, and BGP; support IPv 6 dynamic routing protocols such as RIPng, OSPFv 3, ISISv 6, BGP 4 +; support BFD for OSPF, BGP, IS-IS, Static Route; support IPv 6 ND and PMTU discovery; 6, DC feature: support Vxlan, and support BGP EVPN feature; 7. Security: Support against DOS, arm p attacks and ICMP attacks; support the combination of IP, MAC, port and VLAN binding; 8. Configuration and maintenance: support Telemetry; support SNMP V1 / V2 / V3, Telnet, RMO N, SSH; |
| 3 | Access to the switch | 1. The exchange capacity is 336 Gbps, with the lowest parameters listed on official website; the packet forwarding rate is 108 Mpps, with the lowest parameters listed on official website; 2.24 Gigabit ports, 4 Gigabit SFP +; 4 Gigabit multimode modules; 3. Support MAC address 16K; support ARP table item 4K; support RIP, RIP ng, OSPF, OSPFv 3 routing protocol; support IPv 4 FIB table item 4K; 4. Port-based multicast traffic statistics; support CPU protection; support limit of port reception rate and transmission rate; support ERPS Ethernet protection protocol (G .8032) ； 5. Support for Telemetry technology. |
| 4 | Backup the all-in-one machine | 1. Support mainstream operating system and file system backup, including various operating systems under Windows, Linux and Unix; 2. Support the online backup of mainstream databases and applications under multiple platforms, including: Oracle, SQL Server, MySQL, Exchange, SharePoint, ERP and other applications; 3. There is no need to restore the virtual machine backup data, directly browse and restore the files in the virtual machine, greatly improve the recovery speed and simplify the recovery operation. Support virtualization: VMware, Hyper-V, Citrix Xen, Red Hat Virtualization, Amazon, Azure, Nutanix Acropolis, OracleVM, OpenStack; 4. Provide the particle recovery ability for the files and applications in the virtual machine, and the whole process without any script; 5. Support Oracle, Exchange, mainstream data file system backup operation and recovery, and automatically continue to work from the breakpoint without user intervention; 6. Support file system and related application continuous data replication protection, and can create application consensus point snapshot, less data loss and ensure application consistency, meet the ROP / RTO requirements for user disaster backup, support one-to-many, many-to-one mode. Can support: Windows, AIX, HP-UX, Linux, Solaris, Exchange, SQL, AD, Oracle; 7. For files backed up at the storage snapshot and data block level, they can share access directly by using CIFS / NFS protocol without recovery; 8. Support automatic recovery drill. Through the formulation of recovery drill strategy, automatic and complete recovery drill test, it can be restored to the physical machine, virtual machine, or private cloud and public cloud; 9. Follow the IPMI 2.0, SMBIOS, SAS 2.1, ACPI, and IP protocol standards; 10. The configuration of 50T back-end data backup capacity use license; 11. Hardware configuration: 2 * 900W, 2 * C 4215R (8 Core @3.2GHz), 4 * 32GB of RAM, 2 * 600GB SAS, 12 * 10T SATA, 2 * GE + 2 * 10GE. |
| 5 | Virtualization software | 1, virtualization supports dual architecture deployment, can be directly installed on the physical server based on x 86 architecture or ARM architecture, can benefit the old network x 86 devices, unified management; 2. Support the online or offline adjustment of virtual machine specifications, including CPU, memory, hard disk, network card and other resources, and support the restart coming into effect; 3, the virtual machine supports BIOS and UEFI start mode, and the administrator can customize the start media, such as network start, optical drive start, hard disk start, etc., and can accurately specify the start order, need to provide interface screenshots; 4, X 86 and ARM server deployment, can provide virtual machine basic lifecycle management functions, support deletion, mobile, cloning, migration, VNC login, snapshot, export, restart, close, forced restart, forced close and other operations; 5. Support virtual machine HA, allowing the configuration of the number of HA reserved hosts in the cluster, to ensure that there are enough resources to switch when the virtual machine failure, support the configuration of HA virtual machine or not processing after the storage failure; 6. X 86 scenario supports GPU virtualization, virtual a physical GPU card into multiple vGPUs to meet the requirements of the latest DirectX and OpenGL specifications; 7, support GPU equipment, SSD equipment directly to the virtual machine, the combination of soft and hard to improve the related graphics processing of the virtual machine, storage IO and other high performance requirements; 8. Support setting the thermal migration properties of intergenerational CPU virtual machines in clusters. Support the same CPU vendor different CPU model server set up in the same logical cluster, and support virtual machines between different CPU model servers without interruption of hot migration; 9. Compatible with the mainstream storage array products in the existing market, such as SAN, NAS and iSCSI, brands including EMC, IBM, Huawei, HP, HDS, NetAPP, DELL, etc.; 10. Support the mainstream x 86 and ARM architecture operating system, including Redhat, Ubuntu, CentOS, Winning Kirin, Depth, Fedora, OpenSUSE and other mainstream Linux OS. The bidder shall provide screenshots of the query website and compatibility list; 11. The system supports the management and operation and maintenance mode of "three members", with three roles of system administrator, security administrator and security auditor, to meet the requirements of authority separation in high security scenarios; 12. Support the backup of management data to third-party backup media through FTP, FTP Ps, SCP and other protocols to improve the reliability of management data; 13. Provide a data protection system to realize the backup and recovery of virtual machine snapshots; 14. Provide the remote disaster recovery strategy configuration and execution components to realize the remote disaster recovery protection of the business system; 15. Provide 20 CPU authorization. |
477
被折叠的 条评论
为什么被折叠?
到【灌水乐园】发言
