W H I T E P A P E R
Extending the SAN for SANbox 9000
Core Switches
QLogic Intellegent Storage Routers
Extending the SAN Beyond the Data Center
Executive Summary
A new challenge faces IT professionals, dubbed the corporate
“extrastructure.” This includes devices and technologies outside
the IT department’s direct control, but are vital to the proper
operation of dispersed SANs and data resources. For example,
most corporations rely on third-party vendors for the bandwidth
and connectivity technology outside of the brick-and-mortar
confines. Since implementing a private global WAN would be cost
prohibitive, the company needs to bridge data centers without
sacrificing the benefits of a single-site operation.
The best way to mitigate the limitations of a decentralized data
environment is to create a core-to-core bridging strategy using
standard, proven technologies. Core-to-core bridging is the
QLogic term for linking the core networks between data centers.
Data center switches, such as the SANbox 9200, can access
remote SAN islands with the proper deployment of gateway
routers. The SANbox 6142 Intelligent Storage Router can bridge
core switches across the WAN with fast and efficient Layer 3
routing. The routing can even be easily implemented across core
switching technologies from other vendors, for example Brocade
and Cisco.
Mission critical applications, such as data migration, replication,
and disaster recovery, can now be integrated across multiple
sites for better utilization, higher levels of protection, and faster
recovery times.
Key Findings
The SANbox product family can implement a cost-effective and
efficient data center core-to-core bridging strategy to achieve the
following solutions:
Greater resource utilization in SAN islands
•
Higher data protection levels with off-site migration,
•
replication, and disaster recovery
Better operational efficiencies
•
Faster gateway bridging across homogeneous and
•
heterogeneous SAN fabrics
Easier management of remote data resources
•
Core-to-core fabrics can be bridged without merging SANs
•
Extending the SAN for SANbox 9000 Core Switches
W HI TE P AP ER
Introduction
The corporate infrastructure has evolved over the past few decades
to the point where “infrastructure” no longer provides an adequate
definition. With global offices, remote locations, and a dispersed sales
force, the computing environment now requires an “extrastructure”
to provide data when and where it’s needed. The extrastructure is
much different than an infrastructure because so much of the WAN
connectivity is beyond the control of the IT professional. This paper
looks at best practices for implementing distance connectivity
topology that still maintains the requirements for flexibility, speed
and management available within the data center.
The standard data center architecture consists of servers at the
edge of the Fibre Channel SAN connected to data storage through
distribution and core switches. System administrators can easily
design, control, and manage this environment. However, when they
need to connect multiple data centers, system administrators must
investigate and properly evaluate each one with their own edge-tocore topology, new challenges, and criteria. This paper investigates
the options for creating core-to-core SAN solutions.
An intelligent storage router can provide SAN-over-WAN connectivity
to extend the data center core beyond a single location or campus
environment. This can be done while still maintaining all of the data
center requirements mentioned above. This allows for a unified
business approach to computing needs and data access. Not that
all users in all locations gain unlimited access, but the policies
and procedures required to run a sound business can be extended
beyond the headquarters or main data center to all users throughout
the enterprise.
There are three basic segments to managing the extrastructure:
Capacity allocation approach – Too often, departments and/or
1.
locations demand storage capacity for the wrong reasons. They
“what if” themselves into deploying excess capacity: “What if we
grow faster than expected? What if the requirements were under
estimated? What if we have a surge in peak demand?” A capacity
allocation approach focuses on assigning data storage when and
where it’s needed.
Centralized management approach – An IT department spends
2.
a great deal of time implementing policies and procedures that
guard the company against pitfalls of intrusion, collusion, and
regulation. However, these safeguards are difficult to manage
and enforce across a fractured extrastructure. A centralized
management approach focuses on standard management
policies across the extrastructure.
Distributed network approach – Even though many companies
3.
have basic connectivity between remote offices and sites, the
capabilities of the global network remain fractured. Each office or
data center operates as an independent island within the larger
scheme of data management and resource sharing. A distributed
network approach focuses on high-speed resource sharing
between SAN islands.
SSG-WP07009 2
SN0130966-00 A
Extending the SAN for SANbox 9000 Core Switches
Business Needs within the Extrastructure
Data protection – After keeping applications available to run the
business, the most important function of the corporate IT department
is to secure and protect the data assets of the institution. Growing
data and expanding extrastructure push operating scenarios and
data systems to the limit. The opportunity cost of lost data could be
catastrophic to the company.
Regulatory compliance - Across the globe, changing regulatory
requirements from most governments place greater demands on
businesses. The data outlined by these regulations is often outside
of the data center, but needs to be protected, secured and accessed
globally. US regulations such as SEC 17a-4, Sarbanes Oxley section
404, and HIPAA drive the need for new storage strategies in every
organization. This means that those servers, and the storage that
supports them, need data protection and management capabilities,
often associated with enterprise SANs, traditionally found only in
central IT departments.
Resource limitations – Shrinking IT budgets require that utilization
and leveraging of available resources be maximized. Redundancies
or over capacity from one SAN island to another can’t be tolerated.
Bridging and sharing existing data devices can help ease the burden
on equipment and management.
Evaluation Factors for Building the Extrastructure
The next step is to assess the requirements of the SAN-overWAN application. To help define the requirements of your specific
application, consider the following:
Data Targets – The volume of the target data is fundamental
•
for establishing the performance requirements of the WAN
connections.
Availability Requirements – Does the connection need to be a
•
“best effort” or 24/7? This is key when determining the type of
software, equipment, and WAN connection to implement.
Write Transfer Objective – How much time is available to transfer
•
the data? This is one of the key factors in determining the size
and/or number of telecommunication lines needed.
Recovery Time Objectives (RTO) – How fast does the data have
•
to be recovered? How often do you think you will have to do this
recovery?
W HI TE P AP ER
Budget – Examine not only the cost of the networking equipment,
•
extra disk storage, and required software, but also look at ongoing
WAN line cost to get a full TCO/ROI analysis.
Data Access Model - What kind of data access is required
•
asynchronous or synchronous transfers? If asynchronous
is acceptable, using it can lower the cost and performance
requirements, instead of accommodating peak synchronous
demands.
By answering the above questions, you can systematically eliminate or
include specific solutions and offerings. For example, by determining
that there are varying classes of data being transferred from the SAN
over the WAN, it may be discovered that QoS functionality is required
within your connecting equipment. Knowing the answers to the
above questions, as well as your application availability requirements,
places your IT team in the best possible position to narrow down
choices between solutions.
Issues with FC Core Bridging
The core of the FC fabric is built around high-end director class
switches, such as the QLogic SANbox 9200. This switch provides
solid and proven technology for sustaining storage area networks.
However, expanding the enterprise usually leads to multiple SAN
cores and a patchwork architecture. The investment in the SANbox
9200 should be leveraged for maximum utilization and sharing of SAN
resources. An ideal infrastructure would locate all core devices close
enough to bridge with standard, short-wave Fibre Channel products;
however, this is usually not possible. Although the issues may be
diverse, the solutions are straightforward.
Some of the specific issues include:
Long-distance bridging (>10km) of homogeneous cores – So
•
the IT department was lucky enough to maintain a single-vendor
policy for their FC fabric, but SAN islands have developed, causing
issues with policy management, utilization, and disaster recovery.
Bridging the islands is imperative to maintaining best practices.
Long-distance bridging (>10km) of heterogeneous cores
•
– The worst-case scenario is usually what hits us in real life, and
this situation is no different. Multiple SAN islands have evolved
within the organization, requiring that the SANbox 9200 bridge to
Brocade/McData or Cisco fabrics. Not only do we have to solve the
distance issue, but the compatibility issue as well.
SSG-WP07009 3
SN0130966-00 A
Extending the SAN for SANbox 9000 Core Switches
W HI TE P AP ER
Bridging Topologies
To achieve higher levels of functionality and better management,
vendors usually define proprietary ports, such as E_port, F_Port,
G_port, and TE_port. Within those port types, vendors can have their
own unique classifications and requirements.
If you need to bridge SAN fabrics across vendors, you must often
sacrifice these proprietary features. This requires an intermediary
between one core and the other to allow communication and maintain
neutrality without merging the SANs.
The best way to bridge SANs without merging them, which forces
IT managers to abandon the advanced vendor features, is to use
routing based on N_ports. The N_port definition allows routers or
other SAN devices to log into a SAN as a native device. It doesn’t
cause compatibility issues with the core switches because it does not
merge the fabrics.
A long-distance bridge can be created between SANs with a router
using N_ports. Each core sees an attached native device, but no direct
view of the other switches. SAN management is simplified, without
merging the fabrics. The router can also use Ethernet ports to bridge
across the WAN to the remote SAN island.
The SANbox 6142 provides this type of routing capability. The QLogic
SANbox 6142 Intelligent Storage Router leverages N_ports for
simplified connection to the SANbox 9200 switch and gigabit Ethernet
ports for communication across the LAN or WAN. The basic topology
is shown below.
This network configuration allows SANbox 9200 switches in separate
locations to share data resources across the WAN, without having to
merge the SANs. See the section below on SmartWrite™ technology
for a full description of this technology and the benefits of Layer-3
routing.
SSG-WP07009 4
SN0130966-00 A
Extending the SAN for SANbox 9000 Core Switches
W HI TE P AP ER
Using a combination of core and distribution switching, very costeffective strategies can be deployed across the extrastructure.
Distribution switches, such as the SANbox 5600, provide flexible and
expandable fabric services in locations with medium-duty data traffic.
Therefore, various combinations of the SANbox 9200, 5600, and 6142
can meet the changing needs of the corporation, without an overinvestment in SAN hardware. The basic concept for a multi-tiered
switching topology is shown below.
In this scenario, smaller data centers use distribution switches to
create the SAN fabric. The SANbox 5600 provides 10GB ISL backbone
ports for expansion, without sacrificing switch ports. Meanwhile, the
main data center utilizes the power and cost-efficiency of the core
chassis to handle the major switching needs of the enterprise. As
shown, each location is connected with the SANbox 6142, allowing
for data migration, replication, and disaster recovery, as discussed
further below.
SSG-WP07009 5
SN0130966-00 A
Loading...