Virtual Machine Deployment: Seven Key Challenges we Can’t Ignore

There are seven areas of concern that accompany any major virtual platform implementation
or migration. These critical pain points directly impact two cornerstones of the data center:
network and storage.

1 Challenge: Performance and Availability Issues

Performance and availability issues—often created by a move from physical hardware to
virtual hardware—as well as VM saturation can cause your application networking resources
to be depleted at a much faster, and often unanticipated, rate. I/O intensive operations get
bogged down in the virtualization translation layer. Saturation of the network card and
software switch on the physical host might cause performance issues, reduced bandwidth,
and increased latency.

Solution: Offload Application Resources to a Purpose-Built Appliance
By offloading application networking resources onto a purpose-built appliance, your
organization can conserve bandwidth, reduce latency, and improve the performance of
I/O intensive operations. F5 BIG-IP® Local Traffi c Manager™ (LTM) uses SSL processing,
caching, and compression to dramatically improve application performance while keeping
VM resources focused on what they do best.

2 Challenge: OS Virtualization Fails to Virtualize the Application
Virtual infrastructure platforms include software that can migrate VM instances from
one physical device to another. However, there might be a lapse in the availability if the
connections to the application and application persistence for the user are lost during this
live migration.

Solution: Improve Application Performance with Automated Traffic Monitoring
By monitoring applications running on the VMs, and directing traffic to the appropriate
VM as application availability and response time fluctuate, your organization can improve application response times and availability while maintaining connection persistence.
BIG-IP LTM accomplishes all of this automatically; when application traffic is routed through
BIG-IP LTM, images are added to the pool and configured for the correct application before
they become available.

3 Challenge: Additional, Unanticipated Costs—the Virtual Solution May Cost More
than the Physical Problem

Additional costs often result from implementing OS virtualization; new hardware and
software licenses can be required to solve problems with availability, performance, and
management. What’s more, as VMs burden the existing infrastructure, requirements
grow for application and storage networks. Management of new tools requires increased
headcount and training. And VM sprawl can create orphaned virtual disk images that take
up critical storage space.

Solution: Optimize Existing Resources with Automatic Storage Tiering, VM Load
Balancing, and Virtual Infrastructure Scaling

Planning ahead and implementing strategies that make the most of your existing resources
can ensure a smooth deployment, reduce added headcount, and limit your exposure to
CapEx and OpEx overruns. With BIG-IP LTM already in place, there’s no need to buy new
gear because BIG-IP devices are just as effective optimizing virtual servers as physical
servers. An investment in BIG-IP LTM is also smart planning for future growth because it
helps optimize the virtual infrastructure and scales with it. In addition, using F5 ARX® file
virtualization solutions to virtualize the storage network enables your network to adapt to
VM requirements without re-architecting. Tiered storage with ARX automatically places
current or critical virtual disk images and data fi les on tier 1 and stores less frequently used
files and images on lower tiers—preventing sprawl and reducing backup costs.

4 Challenge: Advanced Virtualization Features Go Untapped

New virtual platforms include many advanced networking technologies, such as software
switching and support for VLAN segmentation. However, these features are localized and
isolated to the VM platforms. They are not integrated with the rest of the Application Delivery
Network because there is no sharing of information directly between VMware or Hyper-V
and the network. And in some VMware Distributed Resource Scheduler (DRS) or VMotion
deployments, the current networking infrastructure might not be able to support live
migration or virtual software switching.

Solution: Integrate the Application Delivery Network and VM Platforms for Policy
and Information Sharing

Direct integration between Application Delivery Networks and VM platforms enables your
organization to get the most out of those platforms while reducing costs and management duties. The F5 deployment guides for VMware ESX/vCenter and Microsoft Hyper-V/System
Center help make BIG-IP LTM hypervisor-aware. They provide instructions for direct
integration between BIG-IP LTM with F5 iControl® API and the VMware vCenter and
Microsoft System Center API, enabling BIG-IP devices to feed networking and application
information to the respective VM management platforms. iControl integration also
enables configuration and policy information to be shared between BIG-IP LTM and
the management platforms, allowing configuration changes to be automatically passed
between the application network and those platforms.

5 Challenge: Managing Explosive and Unexpected Storage Growth While Maximizing Resources

Prior to server virtualization, OS and data files typically reside on internal storage in the
physical server. These files are moved to external shared storage in virtual environments.
OS drives are converted to fl at-file VM disks which consume 10 to 100s of gigabytes
of networked storage each. Little used or inactive virtual disks can remain on expensive
storage well after they are needed, driving up storage costs.

Solution: Move Virtual Disks to Lower-Cost Storage with Automated Storage Tiering

Storage tiering—the process of moving stored assets according to their value—can
dramatically drive down costs and improve resource utilization. Automating the process
takes it a step further, simplifying storage administration and freeing up IT resources. Using
ARX file virtualization solutions for policy-based storage tiering and centralized storage
management enables your “inactive” virtual disks to automatically and non-disruptively
move from expensive tier 1 storage to more cost-effective tier 2 storage. Virtual disks can
also be dynamically moved back to tier 1 storage without impacting users or applications.
Management of large volumes of virtual disks is simplified, enabling multiple, heterogeneous
storage systems to be merged into large, shared pools.

6 Challenge: Congested Storage—Data Pipes Can’t Handle the Volume

OS virtualization can dramatically increase data storage traffic. Passing large amounts of
data from multiple guests through one host storage network connection, such as network
file system (NFS), can cause instant bottlenecks and flooding. Moving large virtual disk
images outside the LAN causes extensive delays and floods the much smaller WAN
connections. And unplanned VM migrations resulting from VM sprawl can bring the
network to a standstill.

Solution: Reduce Data Volume with TCP Optimization, Data Compression, and
Automated Storage Capacity Balancing

By optimizing TCP connections, improving file transfer speeds, and effectively balancing
storage capacity, your organization can reduce bottlenecks, lower bandwidth costs, and
optimize its bandwidth usage. F5 WAN optimization solutions deliver TCP optimization, data compression, and intelligent byte caching—accelerating file transfer speeds by up to
three times and reducing bandwidth utilization. In addition, ARX file virtualization solutions
provide automated storage capacity balancing, ensuring that virtual disks are created on your
highest performing storage resource while balancing virtual disks across multiple storage
resources—eliminating bottlenecks, aggregating capacity, and increasing utilization.

7 Challenge: Management Complexity

Managing VMs as part of the complete management solution can be a struggle. This
includes managing the VMs themselves, as well as managing all parts of the data center
as one unit. The hypervisor and the host system are two new components that are not
part of existing data center management solutions; it is critical to be able to manage these
devices and understand their impact on performance. What’s more, built-in management
tools for VM platforms only manage the virtual resources and do not take into account any
external information.

Solution: Create a True Application Delivery Network

By managing VMs, the application network, and the storage network together, your
organization can dynamically provision resources, simplify administration, and free up IT
resources. F5 BIG-IP and ARX products help manage the virtual infrastructure by focusing
on application and storage networks. The F5 deployment guides for VMware and Microsoft
virtual platforms help make BIG-IP LTM hypervisor-aware, providing direct integration
between F5 solutions with iControl and the VMware vCenter and Microsoft System Center
APIs. This means you can dynamically provision resources by integrating information from
the network and the VMs, which enables your business to create and manage a true
Application Delivery Network focused on your applications.