HP Extended Cluster for RAC White Paper
HP Extended Cluster for RAC
Continuous availability with the flexibility of virtualization
Executive summary............................................................................................................................... 2
Business dilemma ................................................................................................................................3
The solution......................................................................................................................................... 3
Context for the solution ..................................................................................................................... 3
Solution overview............................................................................................................................. 4
Solution components......................................................................................................................... 4
HP Extended Cluster for RAC............................................................................................................. 7
The tests ............................................................................................................................................. 8
Testing methodology ........................................................................................................................ 8
Test architecture............................................................................................................................... 9
Supported hardware configuration ................................................................................................. 9
Supported software configuration................................................................................................. 10
Test descriptions............................................................................................................................. 10
IPC test description ..................................................................................................................... 10
I/O test description .................................................................................................................... 10
OLTP test description................................................................................................................... 10
Failover test description............................................................................................................... 11
Test results and tuning..................................................................................................................... 11
The next steps.................................................................................................................................... 12
Solution flexibility .............................................................................................................................. 12
Conclusions ...................................................................................................................................... 13
References ........................................................................................................................................ 13
For more information.......................................................................................................................... 14
Executive summary
The concept of optimized resource utilization has moved to the forefront of every IT organization’s
agenda. The maturation of the clustering market has forced hardware vendors to differentiate their
products to gain a competitive edge. Consequently, these vendors have begun to offer some variant
of virtualization as the latest trend in high-availability capabilities.
Traditionally, every IT manager was confronted with the dilemma of having to balance risk and cost.
Reducing risk in an IT environment necessarily came at high costs; conversely, lowering one’s costs
engendered substantial risks. The prevailing model for risk reduction focuses on the duplication of
critical computing components, the elimination of single points of failure, and the potential use of a
secondary facility to continue operations in the event of a catastrophic loss of a primary data center.
Although this approach demonstrates diligence toward risk management, it also comes at the expense
1
of duplicative and underutilized resources. HP Extended Cluster for RAC (Real Application Clusters)
now addresses part of this utilization shortfall by taking advantage of the advancement in
virtualization technologies. HP is regarded as the industry leader for virtualization and is continually
raising the bar on optimized resource utilization.
The HP product line encompasses both high-availability and disaster-tolerant solutions. Coupled with
its powerful virtualization capabilities, HP is the only vendor offering products that seamlessly
integrate the three concepts of high availability, disaster tolerance, and virtualization. To illustrate this
compelling value proposition, this white paper will examine the HP solution for Oracle® RAC.
HP Extended Cluster for RAC combines the high-availability product, HP Serviceguard Extension for
RAC (SGeRAC) and the disaster-tolerant solution, HP Extended Campus Cluster. Moreover, when
integrated with BEA WebLogic Server and virtualization tools, the net result is a compelling solution
that specifically addresses risk reduction and capitalizes on IT investment in any environment that uses
an Oracle9i RAC database.
HP Extended Cluster for RAC enables a single Oracle database to be shared across two data centers,
up to a distance of 100 km. Because the two sites are functioning as a single virtual entity, all
resources can be utilized at all times. The ability to meld distributed IT components into a
homogeneous resource pool is made possible by the HP Virtual Server Environment (VSE) portfolio of
solutions.
Virtualization enables full resource utilization within an enterprise while still maintaining previous
levels of availability and data protection. Because data is replicated and synchronized between data
centers, the entire computing environment is said to function as a single virtual entity. A distance of
100 km between the data centers, for instance, will not preclude an administrator from managing the
application and data as if it were a traditional RAC application located in a single data center. Thus,
whether an IT environment is located in one data center or across two or more, HP solutions are
designed to protect mission-critical applications against hardware and software failures, planned or
unplanned.
With the VSE, HP has created a unique solution that provides unrivaled levels of risk management
and return on IT investment, and it has simultaneously enhanced an enterprise’s ability to rapidly
accommodate and capitalize on volatile business conditions.
By extending a RAC solution over distance with more efficient use of resources, HP has demonstrated
that global virtualization has arrived.
1
Oracle RAC technology allows multiple instances of an application to access a single logical database across multiple servers, with all nodes
able to concurrently execute transactions against the same database.
2
Business dilemma
If a group of CIOs was asked to identify their primary cause of stress, the majority would answer
“controlling IT risks and costs.” Indeed, many would say that this is the most important challenge
within a CIO’s charter.
In striving to balance risk management and fiscal responsibilities, every IT manager is faced with a
difficult choice. On one hand, risk can be reduced by investing in duplication and redundancy, but
on the other hand, costs can be driven down by exposing the organization to increased levels of risk.
Therefore, simultaneously lowering both risk and cost ostensibly is mutually incompatible.
A previously acceptable and widely adopted approach to risk management was for an IT
organization to purchase excess equipment to create a pool of surplus CPU, storage, and connectivity
resources. This method was effective for dealing with various adverse situations. However, in today’s
cutthroat and fiscally cautious business climate, it represents a highly visible example of poor
utilization of IT assets.
To allow for changes in business conditions, IT systems are typically configured to handle maximum
anticipated loads and designed to be scalable, with additional hardware. This approach provides a
fair degree of flexibility, but it does so at the expense of asset utilization. HP studies have shown that,
on average, system use ranges from 20% to 40% of available resources, yet it is not unusual to see
many applications bottlenecked by resource constraints.
For many years, HP has been providing solutions that maximize clients’ return on IT investment. By
combining innovative technologies with proven industry-standard solutions, HP now ushers in the next
generation of solutions that increases utilization, manages risk, and makes the most of IT
expenditures.
The solution
Context for the solution
The Adaptive Enterprise is the HP vision for helping customers synchronize business and IT to
capitalize on change. Virtualization enables you to balance two seemingly contradictory concepts—
cost and agility—by pooling and sharing servers, storage, networking, and other infrastructure
devices and allocating them across applications and processes as your business demands them. The
new Adaptive Enterprise has management capabilities to sense changes in business demand and
trigger the dynamic supply of virtualized resources.
HP defines virtualization as managing an IT environment as a single entity by pooling and sharing
resources such that supply automatically meets demand in real time. This holistic approach allows the
entire IT resource pool to be viewed as a single virtual entity that can be allocated—through business
priorities and policies—to dynamically serve the requirements of the enterprise.
In isolation, virtualization over extended distances has the potential to actually increase the burden
placed on IT organizations. However, by combining virtualization with powerful real-time automation
tools, HP has been able to create a compelling suite of solutions that exploit previously underutilized
assets without causing additional IT overhead.
HP has included the core of its intelligent virtualization concept in the VSE portfolio of solutions. Each
of the VSE offerings is fundamentally designed to achieve a common set of objectives:
• Improve return on IT investment
• Provide enhanced functionality to accommodate and exploit business volatility
• Enable elevated levels of IT service to the enterprise
• Present unrivaled choices for the management and mitigation of risk
3
Solution overview
At the center of the “risk versus cost” challenge is the question “How can I increase application
availability and resiliency while also aggressively managing costs?” The concept of using multiple
data centers, with replicated system configurations, to provide fallback capabilities through
redundancy has emerged as the preferred solution. This approach minimizes the impact of the total
loss of one data center, but it achieves this through the gross underutilization of IT assets.
HP has addressed this issue with a unique and compelling blend of optimally utilized high-availability
and disaster-tolerant solutions. HP has retained all the advantages of split data centers and is now
offering solutions that virtually eliminate redundant, underdeployed resources. This includes support
for Oracle RAC, which allows one of the world’s most popular databases to be deployed for high
availability and scalability.
Oracle9i RAC allows multiple instances to access a single logical database across multiple servers,
with all nodes able to concurrently execute transactions against the same database. However, for the
systems to be housed in separate data centers, several other key solution components must be present.
These components include data center connectivity, workload balancing, and a variety of service
managers—for cluster arbitration, business policy enforcement, volume management, partitioning,
mirroring and synchronization, and so on—as well as supporting hardware.
To illustrate the real-world feasibility of hosting a single logical database across multiple discrete data
centers up to 100 km, HP enhanced its SGeRAC solution.
Solution components
The highly pervasive RAC configuration features comprehensive high-availability and scalability
functionality. SGeRAC delivers the ability to perform identical operations across two remotely-located
data centers and gives the configuration increased stability through additional high availability and
disaster tolerance. In addition, it produces efficiency gains through the utilization of all dynamicallyallocated resources.
4
Figure 1. HP Extended Cluster for RAC—100 km
Cache Fusion is a key component of RAC that uses cluster interconnect technologies to facilitate virtual
buffer sharing through the use of direct memory accessing. The Cache Fusion architecture creates a
single virtual cache across all the nodes of the cluster. The resulting single view of all buffer caches
minimizes disk I/O by allowing any database request to be served by any node in the cluster. In
other words, Cache Fusion enables shared access to all the data on disk by all the nodes in the
cluster.
The introduction of a second data center creates many advantages for risk management and
mitigation. However, until virtualization with SGeRAC was available, it also had the potential to
introduce significant levels of asset underutilization and operational complexity.
5
Loading...