Managing a Solaris Backup and Restore Initiative
Sun and the Evolving Datacenter Model
Organizations today are reaching beyond the traditional glass house boundaries to interact directly with customers, partners, suppliers, and employees. In this new extended enterprise model, organizations seek to provide universal access to information and services via business portals on the World Wide Web. This shift is born out of the pressure of global competition in a networked world, and places new availability, manageability and performance requirements on a datacenter. Today's IT priorities include: improving business processes, improving customer data quality, developing better customer relationship management, implementing high availability networks, boosting networking bandwidth, and building e-commerce infrastructures.
In this evolving business paradigm, new data is generated at exponential rates, and plays a more vital and publicly visible role than ever before. Consequently, organizations must place a high priority on safeguarding mission-critical data and ensure it remains continuously available. A reliable, flexible, and highly-scalable network-based backup and restore solution is an essential part of this endeavor.
Sun Microsystems, Inc. supports the new business model with enterprise servers, storage products, networking hardware, and related software and services, including leading backup and restore solutions. For up-to-date information on Sun datacenter initiatives, see http://www.sun.com/datacenter. For information on services available from Sun, including professional consulting services, see http://www.sun.com/service.
Sun offers a line of enterprise servers, from the Sun Enterprise UltraTM 5S up to the mainframe-class Sun EnterpriseTM 10000, also known as the StarfireTM server. At the time of this writing, Sun has shipped well over 1000 Starfire servers, which are now operating in 46 countries worldwide.
FIGURE 1-1 Today, many datacenters are powered entirely by Sun Enterprise servers. Starfire servers are deployed in the datacenter shown above.
High availability is a major component of Sun datacenter initiatives. SunUPTM is a collaborative program between Sun, customers, and third parties to analyze, develop, implement and manage services, infrastructure, and products that improve availability. For additional information, see http://www.sun.com/availability
Genesys is the Sun code name for the platform architecture of datacenter.com, the new datacenter model. Genesys is also a program, consisting of products and services, aimed at helping IT organizations move their datacenters into the dot com era. For additional information about Genesys, see http://www.sun.com/datacenter/genesys.html
Sun has received the attention and enthusiasm of today's software developers. Many of the new Internet-enabled applications are written for the SolarisTM operating environment and optimized for Sun systems. Sun is the platform of choice for enterprise resource planning (ERP), electronic commerce, and database/data warehousing applications, as well as server consolidation initiatives.
The Sun-Netscape Alliance, known as iPlanet (http://www.iplanet.com), also supports the portal computing model with software and servicesas well as content and audience reach through the AOL/Netscape end of the alliance. The Sun-Netscape alliance offers a full line of software products including the Netdynamics_ application server, the highest performance application server on market.
To support the new datacenter model, large organizations must employ a backup and restore solution that can scale massively in a widely distributed environment. To help customers meet this demand, Sun offers OEM versions of the two most popular and powerful enterprise backup and restore tools on the market. They are: VERITAS NetBackup as Sun StorEdge Enterprise NetBackUp_, and Legato Networker_ as Solstice Backup._
There are many other backup and restore products on the market, but these two are among the most scalable and robust products available. Choosing a backup and restore tool is one of the most important decisions an IT manager must make. For guidance on this issue, see "Backup Tool Selection Criteria" on page 6.
This book uses the name NetBackup when referring to the VERITAS NetBackup product.
Modern developments in backup technology require significant processing power and I/O bandwidth. Sun Enterprise servers provide scalable symmetric multi-processing, from one to 64 high-performance UltraSPARC_ II processors, with up to 64 Gbytes of memory, and up to 20 Tbytes of disk storage. The advent of scalable I/O platforms such as these enables a database to be configured for the optimal balance of processing power and I/O bandwidth, enabling online backups to proceed with minimal impact on database performance.
The Sun StorEdge Instant Image and Sun StorEdge Network Data Replicator products are not covered in this book. For information on these products, see the BluePrints book: Business Continuity Planning for Sun Microsystems Technologies.
FIGURE 1-2 The NetBackup GUI supports centralized administration of a backup and restore architecture that may include widely distributed nodes located throughout a global enterprise. The Java_ technology version of the GUI is shown.
Implementing a backup and restore architecture scalable enough to meet current and future business needs, requires broad support for the project. Depending on the organizational structure of the business, a diverse set of people and interests will need to be brought onboard. The cooperation and approval from different VPs, business units, and operational groups may be required.
Some of the groups that could be involved are:
Systems AdministratorsThe persons responsible for planning and implementing backup and restore solutions.
Database AdministratorsIncluding DBAs inside and outside the datacenter.
Application Development GroupsThose responsible for developing internal applications and customizing purchased enterprise applications.
OperationsThose responsible for day-to-day operations of the backup and restore tools.
The system administrators and the database administrators must work together to define responsibilities. The DBAs will want to retain control over the database backup and restore procedures. The system administrators will define the ways that file systems are backed up, including files that are created when databases are exported for backup purposes. These two groups should work together to define the backup and restore architecture, evaluate and test the backup and restore tool that will be used, customize the backup and restore tool, and define the operational lines of responsibility.
The mainframe and client-server system administrators need to work together. It is important that these two groups be integrated as much as possible, rather than separated into distinct realms. Integration can increase trust and cooperation.
Negotiations can be carried out to determine service levels for each business unit. Decisions on such things as how long backed up data must be kept online, which data is critical and which is not, what the backup window will be, and other questions, need to be addressed. (See TechEvolve Corp. Case Study" on page 28, for more information on negotiations.) Service levels need to be set up with the various business units and formalized into service level agreements. The customers of the IT services need to pay for what they are receiving, since service levels imply human, software, and hardware resource costs.
In a rapidly growing enterprise, there may already be an established backup and restore infrastructure and set of procedures, however, the current architecture may no longer be suitable. Several different backup and restore tools may be in use, either purchased or developed in-house. Eventually these tools could reach their limitations in terms of scalability, or may become too unwieldy to handle the growing needs of the enterprise.
The solution is to consolidate under a highly scalable toolset and architecture. All groups using the legacy tools will need to agree on their use to make this consolidation successful. This could involve tradeoffs, therefore, everyone involved should be informed of the overall goals. For example, a backup and restore tool developed in-house might serve some purposes well, and might be easy to customize for certain groups. But the tool may need to be replaced with a standardized tool in the interest of overall efficiency. Everyone involved should understand the larger goals, and be prepared to make some concessions if necessary.
To successfully implement a new backup and restore architecture, a plan should be developed based on guidelines in Chapter 4 "Methodology: Planning a Backup Architecture" on page 63. All parties affected by the transition should be involved in developing, approving, and implementing the plan.