Cisco Cisco MDS 9000 48-Port 8-Gbps Fibre Channel Switching Module White Paper
3 of 8
T
E
C
H
N
O
L
O
G
Y
B
R
I
E
F
Copyright The TANEJA Group, Inc. 2008. All Rights Reserved
87 Elm Street, Suite 900 Hopkinton, MA 01748 Tel: 508-435-5040 Fax: 508-435-1530 www.tanejagroup.com
number of Fibre Channel domain IDs when it
comes to dense VM networks. Large virtual
server
comes to dense VM networks. Large virtual
server
deployments
can
require
an
unmanageable number of Fibre Channel
domain IDs, which currently stand at 16 IDs
per
domain IDs, which currently stand at 16 IDs
per
server.
By
deploying
network
components
like
intelligent
HBAs
to
dynamically assign virtual port IDs to VMs,
no such upper limit exists.
This will enable large virtual environments to
take advantage of higher performance and
bandwidth without suffering from domain ID
limitations. (Or for that matter, single zoning
to support VM migrations. Virtual port IDs
enable administrators to retain the virtual ID
even when moving the VM to a different
zone.)
no such upper limit exists.
This will enable large virtual environments to
take advantage of higher performance and
bandwidth without suffering from domain ID
limitations. (Or for that matter, single zoning
to support VM migrations. Virtual port IDs
enable administrators to retain the virtual ID
even when moving the VM to a different
zone.)
Driver #2: Protecting Quality of
Service (QoS). New developments are
providing
Service (QoS). New developments are
providing
new
levels
of
granular
management
and
troubleshooting
for
virtualized environments. Given the difficulty
in performing granular levels of monitoring
and trending over the virtualized network, it
is very difficult to meaningfully assign
performance and capacity monitoring by
application. When each application runs on
its own VM and that VM can be individually
monitored, then IT will be able to fully
protect services levels in even large
virtualized environments.
And by adding granular, policy-driven
provisioning
in performing granular levels of monitoring
and trending over the virtualized network, it
is very difficult to meaningfully assign
performance and capacity monitoring by
application. When each application runs on
its own VM and that VM can be individually
monitored, then IT will be able to fully
protect services levels in even large
virtualized environments.
And by adding granular, policy-driven
provisioning
IT
can
further
optimize
application service levels on VMs. The larger
the virtual environment grows the trickier it
becomes
the virtual environment grows the trickier it
becomes
to
provision
for
potentially
hundreds of applications. By automatically
setting QoS levels and policies in the
virtualized storage pool, it becomes much
easier to provision space for application-
specific VM data.
Driver #3: Improved security. Security
for the enterprise data center requires
protecting data from accidental or malicious
loss or damage, but not at the cost of lowered
performance or scalability. Appliance-based
security measures work for compact physical
networks but can grow to unmanageable
levels in large virtual deployments. Fabric-
based encryption to protect data at rest and
during VM migrations will safely expand data
integrity on virtual networks. For example,
wire-speed encryption built into a virtual
network will secure fast-moving data during
physical or virtual VM migrations.
These drivers are moving the data center
towards a widespread virtual network.
Physically the data center will not be
drastically different. Fibre Channel will
coexist along with Ethernet as it does now
(although
virtualized storage pool, it becomes much
easier to provision space for application-
specific VM data.
Driver #3: Improved security. Security
for the enterprise data center requires
protecting data from accidental or malicious
loss or damage, but not at the cost of lowered
performance or scalability. Appliance-based
security measures work for compact physical
networks but can grow to unmanageable
levels in large virtual deployments. Fabric-
based encryption to protect data at rest and
during VM migrations will safely expand data
integrity on virtual networks. For example,
wire-speed encryption built into a virtual
network will secure fast-moving data during
physical or virtual VM migrations.
These drivers are moving the data center
towards a widespread virtual network.
Physically the data center will not be
drastically different. Fibre Channel will
coexist along with Ethernet as it does now
(although
FCoE/CEE
will
enable
convergence), there will be multiple servers
and heterogeneous storage resources. The
challenge will be to knit these disparate
resources into a single virtual network,
capable of dynamically assigning and
configuring computing resources to fit
applications amid the changing needs of
business.
This will lead to high utilization rates for
storage, seamless VM migrations and
application changes, and high rates of
encryption
and heterogeneous storage resources. The
challenge will be to knit these disparate
resources into a single virtual network,
capable of dynamically assigning and
configuring computing resources to fit
applications amid the changing needs of
business.
This will lead to high utilization rates for
storage, seamless VM migrations and
application changes, and high rates of
encryption
and
role-based
security.
Management burdens will ease across the
data center as automation and a single
data center as automation and a single