If you receive errors when attempting to view this white paper, please install the latest version of
Adobe Reader.
"Our
storage systems offer unmatched business agility, superior application uptime, simplicity of
management, and breakthrough value for your enterprise."
Source : NetApp (Network Appliance)
How to Use a Network Appliance SAN for Server and Storage Consolidation
Network Appliance is also known as :
Network Appliance Performance,
Network Appliance Definition,
Network Appliance Web,
Network Appliance Products,
Best Value Network Appliance,
Network Appliance Deals,
Discounts Network Appliance,

Network Appliance Software,
Source Network Appliance,
Network Appliance Certification,
Network Security Appliance,
Network Appliance Platforms,
Network Appliance Solutions,
Network Monitoring Appliance,
Network Appliance Systems,
Network Appliance Research,
Custom Network Appliance,
Network Management Appliance,
Network Appliance Benefits,
Network Appliance Data.
1 INTRODUCTION
Many companies are experiencing a rapid proliferation of commodity Intel® and AMD® x86 servers.
These systems are typically equipped with a small amount of internal or direct-attached disk storage.
As server numbers grow into the hundreds, the management complexity for both servers and storage increases
exponentially, putting a strain on limited administrator resources and IT budgets.
Server and storage consolidation using VMware® Infrastructure 3 (VI3) software
and Network Appliance™ storage systems significantly reduces these problems. The benefits of VMware are
uniquely complemented by Network Appliance storage solutions. Using NetApp technology, storage can be
consolidated onto one or a few storage systems accessed across Fibre Channel SANs or regular IP networks.
This dramatically simplifies storage provisioning, increases asset utilization, and
allows a company to take advantage of NetApp data protection and disaster recovery functionality.
VMware allows multiple operating system instances to run on a single physical
machine and does not require changes to existing application software. Server consolidation with
VMware results in significant savings on hardware purchases, because the yearly requirement for new hardware
is reduced dramatically. Hardware savings can be determined by comparing the number of servers purchased per
year with and without virtualization. The difference in the amount of dollars spent on hardware purchases
represents the TCO reduction from using VMware. Dividing this number by the years of server useful life provides
the annual savings. Last but not least is the savings in overall power consumption. VMware Infrastructure 3 helps
conserve energy by utilizing a smaller number of servers compared to non-virtualized server environments.
Server and storage consolidations have demonstrated that it is possible to reduce the number of deployed Intel
and AMD-based servers by 50% or more while achieving the following benefits:
- 60% to 80% utilization rates (up from typical rates of <15% at the server level and <35% at the storage level*)
- Application provisioning time measured in seconds, not days
- Response times for change requests measured in minutes
- Zero-downtime hardware maintenance
* These results are typical, actual results may vary.
This paper describes the real-world experiences of a customer that
reduced its x86 server count by 60% and decreased overall operating costs.
Customer Challenges
- Server proliferation results in rising complexity and costs
- Low asset utilization
- Rising infrastructure costs such as power
- Managing more physical servers
VMware and Network Appliance Value Proposition
- Simplified management for both servers and storage
- Reduced operating costs
- Increased business flexibility
- Advanced data protection and disaster recovery
Deployment Results
- Servers reduced>50%
- Application recovery reduced from hours to minutes
- Centralized backup architecture
- Ability to migrate a live OS
- Ability to pool server resources (CPU, memory, etc.)
- Server management cut 20%
- Tapeless backup for rapid recovery
- Bare metal recovery in 30 minutes
- Disaster recovery for all critical business applications
- Dramatically increased IT flexibility
2 BACKGROUND
Like many organizations, the Intel-based server environment of this worldwide auto
parts distribution company has grown exponentially over the years to support additional
business requirements and applications. Prior to consolidation, the company relied on over 100 servers,
many of them several years old. Supporting these end-of-life servers was one of many hurdles the company faced.
Servers were provisioned every time an additional application was needed. Server provisioning took weeks
and sometimes months, when considering budgets constraints. Storage was both internal and direct attached,
so that no two servers could share that resource, which would have helped to lower the overall cost of ownership.
Additionally, the company faced an exponential rise in management costs for both servers and their data stores.
When several weeks of internal studies against similar companies in the same industry showed that the number of
servers was significantly above average for the number of users supported, the company launched an initiative
to lower operating costs by reducing servers and also to reduce overall storage management.
In 2006 the company selected VMware as an enabling technology to replace physical servers with virtual hardware instances
running on powerful multiprocessor systems (typically two- to four-processor servers).
VI3 is supported with either internal direct-attached storage, network-attached storage (NFS-based),
or SAN-based (FC or iSCSI) disk storage. However, NAS and SAN storage (such as NetApp Network Storage)
enables many high availability (HA), disaster recovery (DR), and quality of service (QoS) features in VI3.
The company chose both NAS and Fibre Channel SAN to allow storage to be consolidated onto the existing NetApp
storage systems. Although building a new Fibre Channel infrastructure can be expensive, the VMware/Network Appliance
server to storage consolidation project server count was cut dramatically and overall expenditures were significantly
reduced. Further reductions were made possible by utilizing NAS for many of applications that required fewer
resources to function. Costs were also offset by the advantages of
consolidated storage and the advanced data management capabilities offered by the NetApp platform.
3 ORIGINAL INFRASTRUCTURE
The company’s storage infrastructure utilized both internal disks and direct-attached
disks via SCSI cables to external disk shelves. Applications included mail servers, Web servers, file servers,
and database servers. Servers were typically one- or two-CPU servers with dual Ethernet ports and dual internal drives,
set as mirrored for redundancy. The network was Gigabit Ethernet. Overall CPU utilization was around 15% to 27%.
Out of the box server deployment took as long as 2 months and as little as 3 weeks, depending on budgets and complexity. Storage provisioning was
accomplished relatively quickly; however, many times an additional server was brought online due to running out of storage space.
Managing the network infrastructure became increasingly complex as more and more servers were brought on line.
The current data center at headquarters could house a maximum of 120 servers, a number projected to be
outgrown in two months. Data replication was 15 to 24 months away. A total of four
employees ran the data center, two server administrators and two storage administrators.
4 NEW INFRASTRUCTURE
Five four-processor Xeon servers and five two-processor servers had been configured with VMware ESX software. These servers were configured to access LUNs on the NetApp system in the primary data center.
LUNs are provisioned in their own qtrees. In accordance with NetApp best practices
(www.netapp.com/news/techontap/3248.pdf), each virtualized server is allocated a qtree in a SAN_vmware volume
on the NetApp storage system. In addition, each VMware LUN is formatted as a VMFS-3 volume. This allows the VMware
Virtual Center to be used to migrate a virtual machine from one VMware ESX host to another while the server is live
(using VMotion™). Additionally, the customer now has localized disaster recovery with the
high-availability (HA) feature in VI3. The HA feature allows virtual machines (VMs) on a single ESX server to
be started on another ESX server within the resource pool (which is managed by Virtual Center). If a server fails
completely, all of its VMs start on another ESX server within minutes. This reduces the overall provisioning of servers
to minutes instead of days. Along with VMotion and HA, the customer also utilized the Distributed Resource Scheduler,
which allows the balancing of virtual machines across a defined set of computing resources such as CPU and memory.
CONFIGURING A NETAPP FIBRE CHANNEL SAN WITH VMWARE
Once the appropriate software is installed, configuring VMware ESX to access NetApp storage via
Fibre Channel Protocol (FCP) and NAS (NFS) is simple. VMware allows the administrator to map LUNs to
the virtual SCSI disks used by each virtual machine instance.
Provisioning LUNs on the NetApp storage systems requires the following steps:
- Decide the size of the LUN, the LUN ID number, and where it will reside on the storage system (use VMware as the host OS type).
- If it is not already present, create an initiator group for the host.
- Run LUN setup on the NetApp storage system via the command line, or use the LUN Wizard tool in NetApp FilerView®.
After the LUN is created, use the VMware administrative interface to map the LUN from the storage system to the specific virtual server.
Figure 1) NetApp Fibre Channel SAN configuration with VMware software.
In sizing the LUN for VMware, it is necessary to increase the size slightly to account for VMware
administrative space. VMware uses this space to accommodate information about the instance, and to store the
memory contents of the virtual server when it is suspended.
The company created LUNs that were 2GB larger than the storage requirement to accommodate VMware.
After the company created its LUNs and virtual servers, the LUNs were mapped to individual
virtual servers and the instances started. A migration tool from VMware called VMware Converter was
used to move data from existing physical servers to VMware virtual machines. After a period of testing and
a brief outage for each, the physical servers were retired or provisioned for testing applications.
6 NETAPP AND VMWARE IMPACT
When the project began, the company relied on more than 100 servers in their main data center.
The environment supported Web servers, mail servers, print and file servers, and database servers.
By simplifying its IT infrastructure with VMware and NetApp storage, the company was able to reduce the
number of deployed physical servers to fewer than 15—a reduction of more than 50%—without eliminating any services.
This effort significantly simplified the company’s IT infrastructure while making it easier
to provide appropriate protection and disaster recovery capabilities for critical applications.
As a result, the headcount directly responsible for storage and server management functions was reduced by 25%.
Consolidating server storage on NetApp storage systems has given the company much
greater control over its data. In the past, server and storage provisioning took as long as 3 months.
Application restore time was reduced from days to minutes by using a two-tiered approach (VMware template
deployments and restoring with NetApp Snapshot™ copies). By utilizing the template feature in VMware Virtual Center,
provisioning VMs took a matter of minutes, including the applications that were configured as a part of the template
creation process. Therefore the customer
decided to have gold configurations based on the multitude of various application types.
For individual file restores, the company has completely eliminated a previous service-level agreement of 4 hours.
The metric is no longer needed because restores occur in minutes.
In addition, the previous tape environment had a much higher risk of losing data, and bare
metal recovery of a failed server could take from 8 to 12 hours for critical servers.
Using the HA feature in the Virtual Center allowed the server and application to restart in minutes.
Improved disaster recovery (DR) is another major advantage. NetApp Snapshot technology and SnapMirror® software
have replaced tape backups with a completely tapeless solution. In the past, the company provided disaster
recovery for only a small part of its overall environment. With NetApp and SnapMirror as well as VMware HA,
the company provides DR for all of its critical
applications and can rapidly add DR capability for any application that requires it.
The company has also found that resource usage is very efficient in its VMware environment.
VMware efficiently shares processor and memory resources between instances of the same OS, so fewer processors
and less memory are needed than originally anticipated, and end users report that application performance has actually
increased noticeably. In addition, the company utilized the ability to pool ESX servers to create resource pools and
carve up CPU as well as memory space
to a granular level (per VM). This is a key feature in VMware Virtual Center 2.0.
VMware enables the company to quickly implement new servers for proofs of concept or new projects.
The company no longer has to budget immediately for new hardware and software; it can simply configure a
new virtual machine with any instance of OS supported by VMware. NetApp and SAN technology complement this capability.
Today, the company makes
changes to its storage environment four or five times a week with no disruption to users.
Network Appliance has unified storage with access via FCP, iSCSI, and NAS (CIFS/NFS).
The company also utilized both NFS and FCP to deploy their environment,
resulting in an overall better ROI and TCO from both the server and storage architecture.
Customer benefits:
- Operational savings in managing, provisioning, and restoring servers
- Environmental savings in power, AC, and data center space
- Reduced hardware costs
- Ability to deploy applications based on resource pools
- Data center DR availability with VMware high availability
- Disaster avoidance with VMware Distributed Resource Scheduler
- Simple management with zero downtime using VMware VMotion
- Lower cost with NFS
7 CONCLUSION
VMware allowed the company to reduce server counts by 50% or more without major application changes, while making it
possible to adapt much more rapidly to support new projects and new applications.
NetApp solutions simplified the company’s overall storage infrastructure,
creating a storage system that adapts rapidly to changing needs while improving data protection
and disaster recovery—often by reducing or eliminating the need for tape-based solutions
The combination of VMware software and NetApp storage solutions offered significant advantages for the company’s
consolidating servers and storage to reduce costs and increase operational efficiency
8 ADDITIONAL RESOURCES
“Network Appliance and VMware ESX Server 3.0 Storage Best Practices,” January 2007.
www.netapp.com/news/techontap/3248.pdftarget
“Network Appliance and VMware ESX Server 3.0: Building a Virtual Infrastructure from Server to Storage,”
October 2006.
www.netapp.com/library/tr/3515.pdf
“Server Consolidation and Containment with VMware Virtual Infrastructure 3.”
www.vmware.com/pdf/server_consolidation.pdf
TABLE OF CONTENTS
- INTRODUCTION
- BACKGROUND
- ORIGINAL INFRASTRUCTURE
- NEW INFRASTRUCTURE
- CONFIGURING A NETAPP FIBRE CHANNEL SAN WITH VMWARE
- NETAPP AND VMWARE IMPACT
- CONCLUSION
- ADDITIONAL RESOURCES