Home News and Information Whitepapers Oracle RAC Proposal Template with New Features
Oracle RAC Proposal Template with New Features  E-mail
User Rating: / 5
PoorBest 
Written by HSYS Staff   
Thursday, 20 September 2007 00:00

We have prepared this Oracle Real Application Clusters template to assist your Oracle RAC proposal to your organization.

Overview

The RAC Proposal Architecture at our Prospective Client has identified a two part need for upgrading its Oracle Database Systems. Firstly, Oracle Corporation has announced obsolescence of certain Oracle Server versions. Secondly, additional capacity is needed in order to provide the same level of service to its Client Users.

This plan identifies the preliminary requirements and proposes a three-phase path to address them.

About Us

We offer the following services:

Phase One: Oracle 10g Migration

This phase will provide the Prospective Client with the most current version of the Oracle software and the means for its care and feeding. Scope will be tightly defined to avoid changing existing operations.

Upgrade

We will install and configure the Oracle Database onto a new set of servers, test and production. The Oracle Cluster Services will be configured to provide redundancy and capacity. Once testing has been completed, the database (Production) will be shut down and its data migrated.

Testing

After an initial migration of the test system, ample time will be provided to the Client Users for testing applications.

Management Tier

We will configure a management tier to manage both the Oracle Database Clusters and ongoing backups of those systems. The management tier will store information about backups in the recovery catalog database.

Basic Monitoring

A set of scripts will be delivered to detect faults and alert staff when critical.

Phase Two: Database Systems Roadmap

In this phase we will define the six to twelve month roadmap for functionality that the Prospective Client will provide its Client Users. Phase two defines a plan. Phase three executes the plan.

Enablers and Controls

A moderate separation between development and production database activities will be defined. The necessary enablers to Client Users will be provided.

Failover

A second database cluster will be defined to protect against data corruption and to provide for disaster recovery.

Database Operations

Additional backup schemes will be introduced to provide for object level recovery, flashback queries, and transaction postponement/resumption.

Architecture

Careful focus upon the needs of users and developers will be incorporated into the roadmap. The database could be re-architected to incorporate these needs.

Phase Three: Service Level Agreement

An operational plan will be implemented to guarantee a level of service to the Client Users and to provide a metric to measure this service. From the systems perspective, this plan defines the care and feeding of Oracle. From the Client Users perspective, this plan represents a guarantee.

Database Roadmap

To implement the components defined in the database roadmap, the Prospective Client will leverage a mixture on-staff and external resources. This strategy will provide redundancy while minimizing costs.

Monitoring and Maintenance

Proactive monitoring will alert primary and secondary resources to database faults. Common faults will be fixed automatically via fix-it scripts attached to the grid management server.

Oracle 10g Migration

More than upgrading the database software, this migration will provide database clusters, storage management, new recovery features, proactive monitoring, and an html management tier.

In addition to de-supporting its older database servers, Oracle Corporation has added many new features that improve and simplify the use and maintenance of its software. In order to realize these improvements, the Prospective Client must upgrade its software.

This Oracle 10g migration will include many new server side features and new funtionality. Database funtionality for applications, however, should be limited to existing functionality only. This limit of scope for this phase will help to guarantee its success. What will be included are the following:

  • Development/Test cluster
  • Production Cluster
  • Backup Tier
  • Management/Administration Tier
  • Basic Proactive Monitoring and Resolution

Hardware Requirements

The Prospective Client's Storage Area Network will provide storage to the Oracle 10g Database clusters. The existing layer three switching should provide ample network capacity and redundancy.

VLAN and Link Aggregation

A virtual LAN (VLAN) shall be configured for the cluster interconnect for Oracle RAC.

To maximize throughput, two pairs of LAN ports shall be used to create two aggregated links: one for the public LAN and one for the Interconnect VLAN.

Automatic Storage Management (ASM)

ASM may be thought of as Oracle’s data driven dynamic volume manager. ASM is the system of choice for storing Oracle data and now supersedes cluster and traditional file systems.

Oracle ASM adheres to the “Stripe and Mirror Everything” (SAME) method. ASM automatically distributes I/O evenly across all disk partitions. ASM also provides for easy management of raw device partitions.

The SAN should provide LUNs in uniform groups, for example, groups of 4, 8, or 16 drives. ASM stripe depth will be set to 1mb, same as the uniform extent size.

Architecture

Real Application Clusters (RAC)

Oracle RAC shall be configured to accommodate both the online transaction processing (OLTP) system and the operational data storage (DSS) reporting system.

If one of the servers in the cluster fails, users are automatically and transparently failed over to the remaining live servers. With Oracle RAC, this failover operation can occur in seconds - effectively masking server failures from the users. Oracle RAC allows the Prospective Client to start small and grow incrementally as its usage demands grow.

Flashback Technology

Flashback Database will augment the current RMAN backup scheme. Flashback may be thought of as a continuous backup that allows row and object level recovery from user error. Phase One shall include these useful features.

Project Plan

Database Roadmap

The database roadmap will clearly define a six to twelve month list of goals for the Prospective Client. The database roadmap shall define a strategy for the Prospective Client in the following two areas:
  • Application Database Architecture
  • Database Operations

Application Database Architecture

Careful attention to the needs of the Client Users will provide insight to useful improvements to the database. Ultimately, sound architecture will provide more improvements than tuning, new features, and operations combined.

Service Oriented Architecture

Web services may be utilized to distribute data to and accept results from remote computing farms. Web services are another way to provide access to the database and are especially useful across wide area networks, where security and interoperability are important.

Application Express (Apex)

Oracle’s Apex product is a database hosted system for developing html delivered applictions. Apex is a graphical development environment that requires little or no programming. The Prospective Client could leverage Apex for reporting or table maintenance. Also, Apex allows administrators and non-programmers to deliver secure html reports rapidly.

Change Control

Basic measures shall be created to divide duties between Development, Test, and Production. For application access, the schema owner should not be used. Most importantly, Client Users shall contribute to the creation of this business process. By owning the change control process, they will see it more as an enabler and less as a constraint.

High Availability

Flashback Technology

  • Database Level: Flashback Database restores the entire database to a point in time.
  • Table Level: Flashback Table restores rows in a set of tables. Flashback drop restores a dropped table or index.
  • Row Level: Flashback Query restores individual rows.

Data Guard

Data Guard offers greater protection, performance, and manageability than other remote storage mirroring solutions. It provides protection from data corruption and site failures.
  • Allows “switchover” operation for planned outages.
  • Allows “failover” operations for unplanned outages.

Rolling Upgrades and Patches

Data Guard provides the means to switchover to a standby instance. The switchover is instantaneous and, once complete, allows the production server to be restarted, reconfigured, or patched.

Oracle Real Application Clusters offers a limited rolling patch capability whereby Operating System and software patches may be applied to an individual node in the cluster, without switching over. Any patch which updates the database is not eligible and must be applied after switching over.

Performance

Separate Reporting from Transaction Processing

A separate reporting instance is a strategy more than a feature of Oracle. There are several ways to provide one; however, using Real Application Clusters minimizes the number of moving parts.

RAC provides a means to set one or more of its nodes to read-only. The Prospective Client can use this feature to provide a reporting instance to direct long running queries. This eliminates the overhead of a traditional data warehouse system and will require no additional storage on the SAN.

Table Partitioning

Patitioning can greatly improve the performance and manageability of the Prospective Client’s largest tables. The partitioning option allows the DBA to divide tables and indexes into smaller segments. Instead of taking an entire database offline, a DBA can work on a partition while the rest of the database is still accessible. Having smaller amounts of data to manage saves time and is easier for the DBA to administer.

Columns may be identified for partitioning off of lists, ranges, or a combination of the two. One option may be to partition on genome + creation date. Creation date may be introduced as a surrogate column that is required. A default value may be specified so this column is never null.

Maximum Availability Architecture

Maximum availability is a strategy to provide the highest level of service with the least amount of annual downtime. This architecture includes multiple layers of redundancy and capacity including:
  • Real Application Clusters
  • Automated Storage Management
  • Flash Recovery Area
  • Data Guard
Maximum availability will be the capstone for database operations at the Prospective Client.

Phase 4: Advantages and ROI

Service Level Agreement

For ongoing maintenance and support of its systems, the Prospective Client has the option of using part-time non-staff resources to maintain its Oracle database systems.

With a single, production database, fractional resources may be a good fit for the Prospective Client. This approach has several advantages.

Cost Effective

Using fractional resources will cost less per year than staffing one junior-level database administrator salary and overhead.

Primary and Secondary Resources

The Prospective Client will have at least two resources to call on for emergencies. A single DBA needs training, can get sick, goes on vacation, etc.

Project Augmentation

Development and administration cost can more easily be attached to projects and grants. Downtime between projects costs nothing.

Experience

For less than the annual cost of a junior-level DBA, a team of experts is available on demand.

References

Oracle® Database Concepts
10g Release 2 (10.2)
Part Number B14220-02

Oracle® Database New Features Guide
10g Release 2 (10.2)
Part Number B14214-01

Oracle® Enterprise Manager Concepts
10g Release 2 (10.2)
Part Number B16241-01

Oracle® Database Lite Getting Started Guide
10g (10.2.0)
Part No. B15918-01

Some of the graphics are the courtesy of sales team at Oracle Corporation.
Last Updated on Thursday, 18 June 2009 17:49
 

Contact Us Now!

Interest
  • We respect your privacy and absolutely will not send you spam.