Reference Architecture

vStart 100 and 200 Reference Architecture
Dell Inc. 7
4 Design Principles
This section covers the design principles, requirements, and solution capabilities incorporated in the
vStart solution architecture.
4.1 No Single Point of Failure
The solution is designed so that there is no single point of failure and redundancy is incorporated into
all mission critical components of the solution. Management applications are not architected with this
level of redundancy because the mission critical workloads will continue to operate in the event of a
management application failure. Network redundancy for the mission critical components is achieved
with redundant network interface controllers (NICs) and redundant switches. NIC teaming for LAN and
MPIO for SAN are used to provide failover across the redundant network interfaces.
For iSCSI storage, redundancy is achieved with redundant NICs, switches, and storage controllers. For
both network and iSCSI traffic, the redundant NICs are selected in such a way that they are mapped
across the rack Network Daughter Card (rNDC) and add-in controllers to avoid any single point of
failure. Hyper-V High Availability (HA) is provided by Windows Server 2008 R2 Failover Clustering. The
solution also includes redundant power supplies connected to separate PDUs.
4.2 Physical Separation of LAN and iSCSI SAN Traffic
Dedicated NICs and switches are provided for iSCSI storage traffic to isolate the storage traffic from
LAN traffic. This helps to minimize network latency for storage I/O operations.
4.3 Logical Separation of multiple LAN traffic types
Virtual Local Area Networks (VLANs) are used to provide security and logical separation of various
traffic types required for virtualization.
4.4 Prerequisites and Datacenter Planning
To support either of the configurations, the following components are required to be present in the
customer environment:
 Active Directory® (AD) Domain Services (AD DS) – An AD DS domain must be available on the
network. The Hyper-V hosts will be joined to an existing or new domain. Cluster Services also
require AD DS. Optional management solutions like SCVMM require AD DS; however, this aspect
is out of scope for this paper. Consult with your Dell Sales and Services representatives for
more details.
 Domain Name Server (DNS) – DNS must be available on the management network.
 Network Time Protocol (NTP) Server - NTP is recommended on the management network.
 Sufficient power to support a vStart 100 or 200 must be present. Detailed power, weight, and
cooling requirements for the datacenter are defined in the vStart 100 and 200 Hyper-V Solution
Specifications document.
 Switch Connectivity – The network architecture supports uplinks into the existing switches in
the datacenter. The uplink recommendations are discussed in Section 5.1, Network
Architecture.