10.7 Putting it all together - mission-critical Exchange

10.7 Putting it all together—mission-critical Exchange

As we come to the end of our project of building mission-critical Exchange deployments, we reach not an end, but a beginning. In a way, I have had the easy task. Yours is the more challenging project. You must glean the key points from this book and other sources and apply them to your own environment. As Microsoft’s latest and greatest, Exchange 2003 is not yet widely deployed. In the years to come, we will both have much to learn about deployment, high availability, disaster recovery, administration and management, and other operational issues for Exchange 2000/2003 and future versions. The goal of this book is to point out the key issues that Exchange administrators will encounter, to discuss the technology foundations and limitations of the environment and supporting elements, and to provide some potential techniques and solutions for overcoming them.

We started our journey by looking at high availability and our need to eliminate as much downtime as possible in our Exchange deployments. As Exchange Server becomes more of a cornerstone platform for Microsoft messaging technologies, the need for more nines of availability will escalate.

We need to be aware of what our organizational availability requirements are and to understand what importance the Exchange deployment has in meeting business and organizational objectives. We need to take Exchange and Windows seriously as mission-critical application platforms to meet these objectives. We can start by analyzing and understanding downtime and outages in our Exchange environments today and the impact those outages have on business.

10.7.1 Downtime

We must look inside of the black box of downtime and understand its components. Whether they be poor management, tools, training, hardware, software, environment, or operational and procedural issues, we need to understand where the most significant downtime charges are occurring. Once they are identified, we can concentrate and focus on these areas and apply sound principals, practices, and technologies to reduce or eliminate them. Don’t underestimate the value of knowing what downtime means in your environment, what it costs you, and the ways in which as Exchange environment can fail. This knowledge is crucial to your ability to plan for contingencies and to put measures in place to mitigate the impact of downtime for your Exchange servers and infrastructure.

10.7.2 Understand the technology

A key step in understanding and the foundation of mission-critical systems is to understand the technology we are working with. This includes not Exchange 2000/2003 alone, but Windows Server and all of the components that Exchange is built on. In Chapter 3, I comprehensively covered the Exchange Server database technology, not because the database engine is the only part of Exchange Server that matters (or the fact that I get excited about silly technological things), but because the database engine and its underlying technologies are at the core of high-availability issues. Issues such as storage allocation and design, disaster recovery, clustering, and proactive management all begin with Exchange’s database technology. While other components and technologies in Exchange Server should not be overlooked in the quest for mission-critical messaging systems, in my opinion, understanding the way Exchange Server stores and manages data is most important to our efforts. Take the time to understand the end-to-end service for Exchange deployments and the underlying technologies that are part of their makeup.

10.7.3 Disaster-recovery technology and best practices

Building on Exchange database technology, we can also add disaster-recovery technology. Understanding how Exchange Server provides backup and restore of data is paramount to maximizing system availability. The degree to which we understand the technology for backup and restore operations in Exchange will determine the success we have in planning, designing, and implementing solid disaster-recovery procedures. Our operations staffs must train tirelessly in this area to ensure that, when real-world disaster recovery crises arise, minimal downtime will be incurred with zero data loss. The key to success is the development of top-notch procedures and best practices. Selecting the best disaster-recovery technology available (for both hardware and software) is also paramount. As you are wading through the sea of vendors, you need to have established selection criteria that will allow you to select technologies that ensure mission-critical system capability at the lowest total cost of ownership. Throwing money at disaster-recovery technology does not guarantee success when procedures, practices, and training are inferior. You can make a difference here by taking the time and investing wisely to establish disaster recovery methods for your Exchange deployment that are best suited to your business requirements.

10.7.4 Clustering technology

Clustering technology will also play a more important role in Exchange 2000/2003 reliability. While clustering features in Exchange 5.5 did not give us much to be excited about, Exchange 2003 promises and delivers much more robust clustering options. The key point here will be how we can leverage an Exchange cluster to provide increased availability and to facilitate server consolidation activities. After all, it is only after we are sure that more feature-rich high-availability options like clustering are available for Exchange Server that we will venture down the path of server consolidation to increase the number of users per server. While the number of users per server may not drastically increase with Exchange 2003, the number of users per cluster certainly brings more possibilities.

While clustering does not promise to heal all that ails your Exchange Server, it does provide additional options. Many large deployments for corporate messaging environments as well as deployments in Internet or application service providers (ISP/ASP) will require clustering as a means to achieve the degrees of scalability and server density required in these environments. For the first time (delivered in Exchange 2000/2003), clustering in Exchange Server provides a high-availability option with an adequate ROI that should not be overlooked when building mission-critical systems.

10.7.5 Server and storage technologies

Once you have invested in understanding the technology and developing sound practices for data management and disaster recovery, it is worthwhile to investigate how certain technologies can further your goals. Server and storage technology is a good place to start. For server hardware, we find that Intel has just about commoditized server architectures. However, there are server vendors that are still making a difference, and it is important to align yourself with one you can count on. As an Exchange administrator, it is important that you understand the underlying server technologies (CPU, memory, I/O bus) upon which you build your Exchange servers. I have also stated that I believe storage technology (primary and secondary storage) is the most important component to mission-critical systems. Not only is storage key to holding, protecting, and managing your data, it is also key to the performance and scalability of an Exchange environment. No single component can have such a drastic impact on these areas as storage technology. In Chapter 8, I discussed server and storage technologies from a foundation technology point of view. These technologies are constantly evolving, and it will be important to stay abreast of new trends in these areas such as SANs, NAS, and their convergence into a networked storage paradigm.

In the long term, we will see the evolution from server-centric storage (host-based) to controller-based storage that is independent of the server platform. The goal is an enterprise storage “utility” in which performance, capacity, disaster recovery, and other storage-management concerns are hidden from the view of application implementers and support staff. SAN technology will deliver this in the near future. Alternative high-availability storage technologies such as data replication, BCVs (cloning and snapshots), “serverless” backup (direct SAN-attached device backup and restore), and clustering are also enabled by SAN technology. All of these technologies can be valuable complements to your tried-and-true disaster-recovery and high-availability techniques. In your search for technology areas that have the most impact on your Exchange Server reliability, don’t forget about your server hardware platform, and remember to start with storage technology as your basis for building mission-critical Exchange servers.

10.7.6 Security

Another often-overlooked facet of high availability is security. When they do not overlook security, many Exchange system managers neglect the close tie it has to maximizing system availability and protecting data. Many times, we look to other departments or groups within IS to provide this vital function for our Exchange deployment. However, in building missioncritical systems, we must take responsibility locally for the security of the Exchange environment. Security in Exchange Server encompasses many issues such as gateway configuration, server administration roles and permissions, locking down services and ports, the PKI, encryption, digital signatures, and so forth. Many times, we just think of security for Exchange as it relates to virus protection and measures for preventing attacks of emailborne viruses that result in data loss and corruption, DoS attacks, or minor inconveniences. As a system manager, you may not have sole responsibility for security in your organization, but you must find a way to ensure that the security issues impacting Exchange are at the forefront of your organization’s security measures. Whether it is protecting Exchange servers acting as SMTP mail gateways, implementing store-based scanning, or using other measures such as antispam, you must play an active role in ensuring that the perimeters of penetration (gateways, servers, and clients) are protected from attack. Based on our experiences over the last several years with numerous e-mail-based virus attacks and the increasing problem of UCE/spam, we are painfully aware of how devastating these attacks can be on system availability. Many times, these attacks do not destroy data, but prevent users from accessing the system. The bottom line is that system availability is impacted. We cannot afford to neglect security when planning and building bulletproof Exchange deployments.

10.7.7 Proactive management

If forced to limit this book to two key topics that would have the most drastic impact on Exchange Server availability, I would probably choose storage technology and proactive management. Since this final chapter has focused on proactive management and its benefits, I won’t harp on the subject.

However, I should reiterate the importance of building your proactive management strategy on business-objective-based SLAs. Without SLAs, it is difficult to understand what your proactive management efforts have achieved. Proactive management has three key components that are mutually inclusive— all three need to be in place to have a sound strategy. These are performance management, configuration management, and fault management. You need to define each of these based on your service-level requirements. You cannot handle these components on your own—you must carefully choose the right tools to help you implement them. The upside is that the market for Exchange application, management tools is rich and competitive, with many vendors desiring to ride the Exchange wave. The right tools do not always come at a high price. Many tools such as System Monitor and Event Log come for free in Windows. In addition, Microsoft has also entered the race with MOM and the Exchange MP for MOM. Take the fee-based and free tools and build a management environment that is uniquely suited to your Exchange environment and that provides maximum reliability through proactive management.

Armed with these concepts and tools, you will hopefully gain from this book with some insights and knowledge that will aid you in your quest for mission-critical Exchange deployments. The next steps are up to you. Exchange Server is a key Microsoft technology that is one of the cornerstones of Microsoft’s .NET Enterprise Server family. Success here and now with Exchange as a messaging platform will ensure that you can build mission-critical systems for messaging and collaboration solutions for your organization. Take heart. We are pioneers, and mission-critical Exchange deployments that provide businesses with business value and competitive advantages are within our grasp!