IT professionals attempting to build robust, high-performance clustered systems with Microsoft technology desperately need authoritative guidance written from the perspective of real, hands-on implementers. Windows 2000 Clustering and Load Balancing Handbook has what they've been searching for. Simple enough for newcomers, thorough enough for those experienced with Windows clusters, it's full of practical techniques that make it invaluable in any clustered environment. Joseph M. Lamb systematically addresses each key Windows 2000 clustering technology, including Failover Clustering, Load-Balance Clustering, Distributed Systems Clustering, and Component Load Balancing. Lamb shows how to use clustering to improve fault tolerance and performance, how to design clusters that will operate within switched Ethernet network environments; and how to integrate cluster servers with network services such as DNS, WINS, and Active Directory. The book includes real-world cluster system design examples that show how to maximize both flexibility and scalability, and provides detailed step-by-step instructions for clustering SQL Server 2000, Exchange Server 2000, Windows 2000 file/print shares, Internet Information Server and more.
1. Climbing the Mountain: 24/7 Data Access.
Fault Tolerance. Performance. Scalability. Reliability. Designing Reliable Data Access. Summary.
Cluster History. Clustering for Fault Tolerance. Cluster Types. Cluster Models. Windows 2000 Clustering. Summary.
Cluster Server Planning. Hardware Requirements. Software Requirements. Placement Considerations. Network Services. Design Samples. Summary.
Resource Configuration. Hardware Configuration. Operating System Install. Network Configuration. -Configuring the Shared Disk Array. Clustering Software Install. Testing the Installation. Managing the Failover Cluster.
How Does It Work? Preplanning. Operating System Install. Configuring NLB. -Configuring Internet Information Server. Testing the Installation. -Managing a Network Load-Balanced Cluster. Summary.
Product Overview. Product Features. Installing AC 2000. -Component Load Balancing: Three-Tiered Clusters. Managing AC 2000.
Overview of SQL Clustering. Designing an SQL Cluster. -Installing SQL Server 2000 on a Cluster. -Managing an SQL Server Cluster. Summary.
-Overview of Exchange Server Clustering. -Designing an Exchange Server 2000 Cluster. -Installing Exchange Server 2000. -Managing Exchange Server 2000.
IIS Overview. Clustering the WWW Service. Clustering the FTP Service. Clustering the SMTP Service. Summary.
Introduction. Clustering File Shares. Clustering DFS. Clustering Print Shares. Summary.
So, it is 11:30 a.m., time for lunch. You close the spreadsheet you are working on, lock your workstation, grab your keys, and hit the door. You think that life could not be any better. Short hours, great job as a senior systems administrator, and all the respect in the world from vice presidents who don't have a clue what you do for a living. Ten minutes into lunch, your pager beeps, catching you off guard. Knowing that your pager has not gone off in the three months that you have been at this company compels you to quickly view the message. You look down at your alphanumeric pager to see the message, "Web site is down again...where are you?" Knowing that you are in charge of the Web site, you quickly wolf down your lunch, grab your barely opened soda, and scoot back to the office. Upon entering the office, your boss screams from his office, "Johnson, get in here!" For the next 30 minutes, you receive a thorough thrashing from your superior. It seems that prior to your employment, three other systems administrators had been fired because they could not keep the Web site, the company's main source of income, online. You inform your boss that the problems with the Web site are due to the shabby work your predecessor engineered (an excuse you will be able to use only once) and promise him that it will never happen again. He believes your barrage of techy jargon and says that if you value your job, you had better make sure the Web site is re-engineered in a more reliable fashion. You leave his office a bit distraught, because you don't have a clue how to keep the Web server from crashing again.
This is a typical scenario. Although maybe a bit more dramatic than most, the story is a common one. The rise of dotcoms through the 1990s introduced the world to the joys of electronic commerce. Buying and selling goods online just might be the future of all commerce. However, just as brick-and-mortar businesses cannot survive if they keep sporadic operating hours, companies that build their business around an e-commerce model, but fail to engineer fault-tolerant and reliable applications, are bound to fail. Downtime for an e-commerce business means lost revenue. Downtime for any company is intolerable, but for an e-commerce company, downtime can mean bankruptcy.
How do you build a server that will provide 100 percent uptime? How do you engineer a system that will deliver high performance and high availability? How do you provide data to users in a way that is consistent and reliable? Many innovators have attempted to build computers that employ fault tolerance in every area, including hard disks, memory, power, and networking, but these machines still fall short of the goal. Other attempts have been made to provide redundancy within different types of architectures (such as VAX and mainframe environments), but they have been limited in nature and not a strong fit for e-commerce client/server computing.
Clustering to the rescue! Clustering is a concept that was developed to provide a fault-tolerant design to client/server environments. A cluster is a group of computers that cooperate as one unit to serve a particular application or service. Clustering rescues system administrators and designers by providing the ability to join several machines together to provide the same service. Clusters employ a concept known as failover to ensure that if one server within a group fails, another server within the group will take over the client load.
The early clustering deployments used hardware devices as a mechanism to forward requests for the same IP address to multiple machines, but these solutions were difficult to implement and cost inhibitive. As the clustering technologies have grown, we have seen the clustering market explode. Companies are now deploying fault-tolerant SAN (storage area networks) to host applications and services that used to be hosted on single servers. Clients can now transparently connect to clusters of servers and reliably run applications.
Not long ago, Microsoft decided to join the party by releasing Microsoft Clustering Services (MSCS) for Windows NT 4.0 Enterprise edition. This began Microsoft's journey into the world of clustering. Microsoft now has a suite of products designed to allow administrators the tools and services necessary to deploy high availability Web sites, databases, and applications in a way that distances itself from prior service deployment methods. Through the use of MSCS, servers can be clustered together to provide high availability and reliability. The Microsoft Network Load Balancing (NLB) service allows administrators to build large server farmsclusters of servers that operate together to perform a single function, distributing load across the cluster and providing failover for server failures. Application Center 2000 enhances the abilities of the NLB service by granting administrators the power of a single console where NLB clusters can be created and managed.
These tools, in conjunction with other Microsoft technologies such as Active Directory, Distributed file system, and Internet Information Server, provide a solid foundation for the deployment of applications and services that outperform any system to date. Performance, high availability, and reliability that were once only a distant vision can now be obtained and utilized to deploy mission-critical applications and services.
This book seeks to guide you through the development and deployment of clustered server solutions. The object is not to teach every possible configuration option associated with each technology (for that would be a very large book), but to guide you through the implementation of several different clustering technologies. This book can be used as a handbook for the step-by-step construction of a clustered solution. It discusses not only the functionality of each product, but wise design goals that should be the focus of your implementation. With this book, you will learn
You can read this book from beginning to end, or you can use it as a guide for the deployment of a particular cluster server solution. If you choose to use it as a guide, make sure to read chapter 1 through 3 as a preface to your implementation so that you will have the design skills necessary to implement your solution.
If you are a student learning to construct a cluster server or you work within an organization that has a problem with your building a Microsoft cluster server on production hardware, you are going to need a few things to complete the lessons.
If you want to build a failover cluster, you will need the following items:
If you want to build a network load-balanced cluster, you will need the following items:
If you would like more information about clustering, need help with your clustering solution, or just want to submit feedback, feel free to submit your responses to
firstname.lastname@example.org or visit the Web site of Qsource Networks at