Due to the confidentiality and sensitivity of data, the isolation of services from one another is very important within the data center. Within the data center, data access needs to be strictly controlled and the business and management networks must be isolated from each other. The management network segment communicates with the tenant network at Layer 3. Tenants access the Portal interface of the storage system through the management network segment and issue management commands such as add, delete, check, and check. The service network segment is responsible for the transmission of service data when the storage space is rolled. The form is attached to the front-end business system through the service network segment and provides services on this network segment.
According to the distributed storage paradigm, the management and services of the storage system belong to two network segments, which are independent and do not affect each other. Data transmission is performed only on the service network segment. Management and services communicate through the server and cannot be accessed through the network.
On the service network segment, each server is planned to be connected to two switches by two network cables. On the management network segment, each server is planned to be connected to two switches by two network cables. The high reliability of the network is provided by the node-level dual-NIC master/slave and the cluster-level switch master/slave. The two network segments use separate physical network cards for isolation and use different VLAN isolation if the conditions are not met.
According to the barrel effect, the overall performance limit of a system is often determined by weak links in the system. When the cluster adopts a hybrid storage configuration, the standard 10Gbps high-speed network can meet the pressure of a fairly large cluster when load balancing and data reconstruction; however, when the cluster adopts the all-flash architecture, the hard disk performance will be greatly improved, at this time the standard A 10Gbps network may become a short board in the system. The 56 Gbps InfiniBand network and even the higher speed 100 Gbps network, which is almost non-blocking communication, breaks through the bottleneck of the internal switching of the storage system. In the InfiniBand network, the communication delay is controlled at the nanosecond level, and the storage information is transmitted in a timely manner. In conjunction with the high speed read and write of the SSD, the communication performance is considerable.
In the process of designing and building distributed storage clusters, the cluster network is a bottleneck. Because the dependence of distributed storage on network bandwidth is very high, all data exchanges need to be conducted over the network, all requiring a high-speed and reliable network environment.
The specific plan is as follows:
1, try to use 10G network interface, and use optical port to connect. The uplink uses 40G interfaces.
2. The network equipment should be as redundant as possible. At least two 10 Gigabit optical ports should be configured to access the switch.
3. In addition to the need to exchange large amounts of data, distributed storage clusters may also perform virtual machine replication and synchronization activities on the network, depending on the number of virtual machines hosted in the system and the number of effective operations. If only Gigabit networks are available at this time, it will be overwhelmed, especially during virtual machine rebuilds and synchronization operations.
4. Put various traffic types (distributed storage networks, management networks, virtual machine migration networks, virtual machine production networks, etc.) in separate VLANs and use the share as a quality of service (QoS) mechanism so that it may appear The desired performance level is maintained in the contention scenario.
5. Dividing different VLANs can also ensure that the distributed storage cluster network is not affected because the IP address conflicts in the distributed storage cluster network will cause the entire distributed storage cluster to become unavailable.
6. To achieve optimal security and performance, the distributed storage cluster network traffic should be isolated to its own Layer 2 network segment.
7. Bind the network card to configure bandwidth aggregation as a measure of availability and redundancy.
The network path Fully redundant virtual network layer avoids service interruption caused by a single network card failure by adopting techniques such as multi-network card binding.
The communication plane of the network sub-plane communication system is divided into a service plane, a storage plane, and a management plane. In order to ensure the reliability of various network plane data, VLANs are used for isolation between different planes. A single plane failure does not affect the normal operation of the remaining two planes. Service plane: It is mainly the communication plane of the virtual network virtual network card, which provides service applications to the outside world. Storage plane: Provides the communication plane for iSCSI storage and provides storage resources for virtual machines. However, it does not directly communicate with virtual machines but is converted through the virtualization platform. Management plane: Responsible for communications such as management, service deployment, and system loading of the entire cloud computing system. - Network card load sharing adopts dual network adapters for each communication plane (service, storage, and management). The dual network adapter adopts the Bonding mode. After the two network adapters are bound to logical “one network adapter,†the synchronization works together. It can not only load balance the access traffic of the server, but also ensure that one of the servers fails. The other network card immediately takes over the entire load. The process is seamless and the service is not interrupted.
MT6-Subminiature Sealed Micro Switch
Features
â—† Designed For Water and Dust Tight(IP67)
â—† Small Compact Sizeâ—† UL&ENEC&CQC Safety Approvals
â—† Long life & high reliability
â—† Variety of Levers
â—† Wide Range of wiring Terminals
â—† Wide used in Automotive Electronics,Appliance and Industrial Control etc.
â—† Customized Designs
Safety Micro Switch,Central Locking Switch,Sealed Waterproof Micro Switch,Subminiature Sealed Micro Switch
Ningbo Jialin Electronics Co.,Ltd , https://www.donghai-switch.com