Deploying Virtual Servers in a SAN Environment By Mark Jones published: Thursday, June 05 2008
IT organizations are turning more and more to server virtualization
as a means to transform data centers into "service-centric" shared pools of
resources which can be dynamically aggregated, tiered, provisioned and accessed
through an intelligent network. Virtualization of standardized server resources
dramatically increases performance levels and reduces total cost of ownership
(TCO), while allowing IT organizations to rapidly deploy and scale resources
on-demand to match business and application requirements.
A shared storage infrastructure, most commonly a Storage
Area Network (SAN), is required to implement the most compelling features of popular
virtualized server environments, such as VMware VI3, including VMotion,
Distributed Resource Scheduler (DRS), High Availability (HA), Consolidated
Backup and ESX Server-remote boot. Consolidating the virtual servers on one (or
more) networked storage arrays provides opportunities for cost-effective and
simplified disaster recovery and business continuity.
Virtual Server
Connectivity
The portability and recovery capabilities of server virtualization
implementations rely on external shared storage, and are most effective in a
SAN connectivity environment. Virtual servers typically reside in the main data
center and draw on enterprise storage resources where the Fibre Channel protocol
dominates. The high performance delivered by the Fibre Channel protocol is best
positioned to serve the higher I/O requirements for multiple virtual machines
running on a single server. SAN connectivity helps enable server
virtualization, while server virtualization drives an increased need for SAN
connectivity.
A major challenge for virtualized server storage
administrators has been the use of the physical Worldwide Port Name (WWPN) of
the Fibre Channel host bus adapter (HBA) to define fabric zones, mask storage
LUNs and configure virtual machines. In addition, virtualized server administrators
have typically defined one zone where all disks are exposed to every virtualized
server to support virtual machine migration to new servers. Such a design
creates concerns for safe isolation of Raw Device Mapping (RDM) disks and
requires the reconfiguration of the network if more than one zone is defined.
The creation of virtual HBA ports (VPorts) using N-Port ID Virtualization (NPIV)
allows virtualized server administrators to bind virtual machines to storage
and define multiple zones using VPort parameters, which creates an easier to
manage and more protected virtualized server environment.
NPIV Overview
NPIV is an industry-standard that extends virtualization to
the HBA by providing a way to assign multiple WWPNs on the same physical link. NPIV
technology virtualizes the physical HBA port configured in a point-to-point SAN
topology. Best-in-class virtual HBA technology allows a single physical Fibre
Channel HBA port to function as multiple logical ports, each with its own
separate identity.
Now, each virtual machine can attach to its own VPort, which
consists of the combination of a distinct Worldwide Node Name (WWNN) and up to
four WWPNs, as shown in Figure 1. Storage administrators who deploy virtual
machines using popular server virtualization environments such as VMware ESX
Server 3.5, and RDM, can create virtual machines that are easier to manage and
maintain. Virtualized servers utilize NPIV to generate a unique VPort, which
each virtual machine can be persistently bound to and which the HBA transparently
registers with the Fibre Channel SAN fabric.
Virtual machine-specific boot occurs seamlessly using the
defined VPort. During virtual machine migration to a new physical server,
storage administrators no longer have to reconfigure their network settings
(i.e. zoning, masking, binding) since they are maintained in the logical port configuration.
Figure 1-Virtual Port Attachment using NPIV
NPIV Use Cases
IT managers are deploying NPIV within virtualized server
environments to enhance storage management capabilities. NPIV is most valuable
in managing storage access for mission-critical, or SLA-driven virtualized
environments, as opposed to the consolidation of less critical file and print
servers, or development and test environments. Below are some specific use
cases that are now possible with an NPIV-enabled virtualized server deployment.
o I/O
throughput, storage traffic and utilization can be tracked to the virtual
machine level via the WWPN, allowing for application or user-level chargeback.
As each NPIV entity is seen uniquely on the SAN, it is possible to track the
individual SAN usage of a virtual server. Prior to NPIV, the SAN and virtualized
server could only see the aggregate usage of the physical Fibre Channel port by
all of the virtual machines running on that server, except for some
vendor-specific LUN-based tools.
o Virtual
machines can be associated to devices mapped under RDM to allow for LUN
tracking and customization based on application needs. SAN tools tracking WWPNs
can report a virtual machine's specific performance or diagnostic data. As each
NPIV entity is seen uniquely on the SAN, switch and array-side reporting tools can
report diagnostic and performance-related data on a virtual machine basis.
o Bi-directional
association of storage with virtual machines gives SAN administrators an
enhanced ability to both trace from a virtual machine to an RDM and trace back
from an RDM to a virtual machine (significantly enhanced with NPIV support).
o Storage
provisioning for virtual machines can use the same methods, tools and expertise
in place for physical servers. As the virtual machine is once again uniquely
related to a WWPN, traditional methods of zoning and LUN masking can continue
to be used, enabling unified administration of virtualized and non-virtualized
servers. Fabric zones can restrict target visibility to selected applications
hosted by virtual machines. Configurations which required unique physical adapters
based on an application can now be remapped onto unique NPIV instances on the virtualized
server.
o Storage
administrators can configure Inter Virtual SAN Routing (IVR) in virtualized server
environments up to the individual virtual machine, enabling large end users to reconfigure
their fabrics, aggregating islands of storage, fragmenting massive SANs into
smaller, more manageable ones and assigning resources on a logical basis.
o Virtual
machine migration supports the preservation of the VPort ID when the virtual
machine is moved to the new virtualized server. This improves the tracking of
the RDMs to virtual machines. Access to storage can be restricted to a group of
virtualized servers (clusters) on which the virtual machine can be run or
migrated to. If the virtual machine is moved to a new virtualized server, no
changes in SAN configuration would be required to adjust for the use of
different physical Fibre Channel ports.
o HBA
upgrades, expansion and replacement are now seamless. As the physical HBA WWPNs
are no longer the entities upon which the SAN zoning and LUN masking are based,
the physical adapters can be replaced or upgraded without changing the SAN
configuration.
Benefits of NPIV for
Server Virtualization
Virtualized server environments provide enterprise data
centers with NPIV support enabled by best-in-class Virtual HBA technology. Data
centers choosing to deploy virtualized server environments with NPIV can
achieve:
o Lower
Total Cost of Ownership (TCO) - Server consolidation through server
virtualization lowers TCO by improving asset utilization and simplifying
management. When used in conjunction with Fibre Channel and NPIV-enabled HBAs,
a single intelligent HBA port can relay the traffic for multiple virtual
machines, offloading network processing, thereby allowing more cost-effective
servers to be deployed.
o Guaranteed
Quality of Service (QoS) - When used in conjunction with fabric QoS, each virtual
machine can be allocated its own logical HBA port, which creates multiple I/O
paths for traffic prioritization.
o Higher
Availability - Multiple logical ports create redundant paths to virtual
machines and their data. They also facilitate the use of standard storage and
fabric diagnostic tools for isolating and resolving issues.
o Role-based
Management and Security - Each virtual machine and its associated storage are
completely isolated from other virtual machines, under control of the
administrator in charge of protecting corporate data.
o Simplified
Management - Eliminates the need to reconfigure fabric zoning and LUN masking
parameters during a VMotion migration.
Related Links:
Emulex ,
Emulex Expands SAN Connectivity Offerings for HP Integrity Servers
Mark Jones
Director of Technical
Marketing, Emulex Corporation
Mark Jones joined Emulex in 2002 and serves as director of
technical marketing. Mr. Jones has over 20 years of experience in the
enterprise server computing industry. Prior to joining Emulex, Mark Jones was
product manager of server systems for Unisys. Mr. Jones holds a bachelor of
science degree in computer science.
|