BladeSystem - General
1752725 Members
5694 Online
108789 Solutions
New Discussion

virtual connect throughput

 
SOLVED
Go to solution
v-zand
Occasional Contributor

virtual connect throughput

Hi everybody recently I'm working on virtual connect and im wondering how 16 blade460 on c7000 with several 10g ports that susceptible to generate a lot of workload in regarding the quantity of uplink on paired vc 40g on bay 1 and 2 which is normally configure in DCs my question is : 1-can 40G uplink even in LACP mode can handle the traffic without bottleneck ? 2-what c7000 is solution for? 3-should i consider 4 or 6 or 8 vc on other bays to rectify bottleneck thank u in advance
1 REPLY 1
AmRa
HPE Pro
Solution

Re: virtual connect throughput

1-Can 40G uplink even in LACP mode can handle the traffic without bottleneck ?

>>It depends on type of network traffic the enclosure configured.  For example, will there be much server to server traffic needed within the enclosure, or is the traffic flow mainly in/out bound of the enclosure.

Virtual Connect works seamlessly with your external network:

  • Uses internal loop prevention algorithms to automatically detect and prevent loops inside a Virtual Connect domain. Virtual Connect ensures that there is only one active uplink for any single network at one time, unless configured for Link Aggregation (LACP).
  • Allows aggregation of uplinks to data centre networks (using LACP and fail-over)

 

Please refer Scenario 10 – Shared Uplink Set with Active/Active Uplinks and 802.3ad (LACP) – FlexFabric-20/40 F8 - Windows 2012 R2 Hyper-V in HP Virtual Connect FlexFabric Cookbook (Page number 223)

https://support.hpe.com/hpesc/public/docDisplay?docId=c02616817

 

2-what c7000 is solution for?

>>Extending the foundation for Converged Infrastructure.

The BladeSystem forms part of the HPE Converged Systems platform, which use a common converged infrastructure architecture for server, storage, and networking products. Designed for enterprise installations of 100 to more than 1,000 Virtual machines, the HP ConvergedSystem 700 is configured with BladeSystem servers. When managing a software-defined data center, a System administrator can perform automated lifecycle management for BladeSystems using HPE OneView for converged infrastructure management.

The HPE BladeSystem c7000 Enclosure goes beyond just Blade servers. It consolidates server, storage, networking and power management into a single solution that can be managed as a unified environment.

With demanding workloads, the increased power supply wattage and mid-plane bandwidth aligned with Intelligent Infrastructure technologies such as Platinum Power Supplies, Intelligent Power Module, and Location Discovery Services have enhanced the foundation for converged infrastructure.

HPE OneView combines server, storage, and networking with control of your data center environment into a single, integrated management platform architected to deliver lifecycle management for the complete Converged Infrastructure.

With Onboard Administrator, iLO remote management, and HPE OneView you can manage your servers and take complete control regardless of the state of the server operating system.

3-Should I consider 4 or 6 or 8 VC on other bays to rectify bottleneck.

>>If Bay 3, 5 and 7 are already populated and you are planning to install VC module in bay 4 or 6 or 8 then please refer below statement.

In horizontal stacking mode, each horizontal bay pair is a separate logical interconnect.  For example, if bay 1 and bay 2 are populated, they form a Logical Interconnect and if Bays 3 and 4 had Ethernet modules installed, they would form an additional Logical Interconnect, which would provide “AIR GAP” between the two “slices” .

In an active/standby network design, all servers would have both NICs connected to the same Virtual Connect network.  All communications between servers within the Virtual Connect Domain would occur through this network, no matter which network adapter is active. For example , if Windows Host 1 is active on NIC 1 and Windows Host 2 is active on NIC 2, the communications between servers will cross the internal stacking links.  For external communications, all servers in the enclosure will use the Active uplink (currently) connected to Bay 1, no matter which NIC they are active on.

Hence when choosing which Virtual Connect network design to use (Active/Active (A/A) vs. Active/Standby (A/S) uplinks), consider the type of network traffic this enclosure will need to support.  For example, will there be much server to server traffic needed within the enclosure, or is the traffic flow mainly in/out bound of the enclosure.

For example; if the solution being implemented will have a high level of in/out or North/South traffic flow, an A/A network design would likely be the better solution as it would enable all connected uplinks.  However, if a greater level of network traffic is between systems within the same enclosure/VLAN, such as a multi-tiered application, then a better design may be A/S, as this would minimize or eliminate any server to server communications from leaving the enclosure to only return on a different uplink/path.

Please refer HP Virtual Connect FlexFabric Cookbook for more information on active/active and active/standby configuration (page number 31)

https://support.hpe.com/hpesc/public/docDisplay?docId=c02616817

I am an HPE Employee.
[Any personal opinions expressed are mine, and not official statements on behalf of Hewlett Packard Enterprise]

Accept or Kudo