Technology provides all kinds of interesting ways to keep servers or the services they provide continuously available, even in the face of failures at many levels. Unfortunately, high availability is a subject that's overstuffed with arcane gadgetry and configurations. It involves powerful and expensive hardware and software options and requires an understanding of engineering tradeoffs at a fairly deep level.
Organizations that excel in mounting and managing highly available servers generally employ highly knowledgeable staff members, many of whom have been working for years with systems and technologies geared to high availability.
Other companies have, basically, two choices: They can train in-house staff. Or they can bring in outside help to design highly available systems and see to their care and feeding over time.
Bone up on the topic
Simply put, developing an in-house IT staff to manage highly available servers means training people to apply the right tools and technologies to make platforms more resilient in the face of problems and better able to transfer control of systems to a backup or fail-safe environment, should that be necessary.
There is no shortage of white papers, articles, tutorials, how-tos and troubleshooting pieces from vendors about highly available servers. These suggestions will jump-start the process:
- IBM offers all kinds of useful product data, general information,
- classes, exams, tutorials, books and even a Recommended reading list: WebSphere Application Server.
- Digging into Microsoft's Web site turns up hundreds of good results, including coverage of database servers, clustering options, BizTalk Server, FAQs, planning guides, resource lists, Microsoft Certified Professionals exam information and course descriptions. I strongly recommend two TechNet articles: "High-Availability System Architecture" and "High Availability and Scalability Planning Process."
- Browsing at Hewlett-Packard Co. turns up product data, certification exams, general discussions of the topic, developer tools, technology descriptions and so on. Two good jumping off links are "HP OpenVMS Systems - High Availability/Disaster Tolerance" and "HP's high availability solutions for the technical data center."
There are also lots of books on this subject, as a quick search on "server availability" will attest. Consultant Mike Tarrani's reading list at Amazon.com, "Guidelines for High Availability," is particularly helpful and informative. He recommends a mix of general titles as well as more focused books by experts at such companies as Cisco Systems Inc. and Oracle Corp., plus other Windows and HP/UX experts.
Vendor credentials focus on specialties
Beyond the reams of books, articles and online content, many vendors offer high-availability training and credentials. If you have specific platform allegiances with IBM, HP, Microsoft or Dell Inc., all of these companies offer technical training courses with availability content. Some even make availability the primary focus.
Microsoft and Dell, for instance, offer IT certification programs that include one or more exams with an availability focus. Microsoft's programs include the Microsoft Certified Systems Engineer (MCSE); and Dell's programs include the Dell Certified Server Professional (DCSP) and Dell Certified Storage Networking Professional (DCSNP) credentials.
HP and IBM offer credentials that focus entirely on availability. At HP, this includes HP Master Accredited Systems Engineer (Master ASE) on Proliant High Availability and Clustering Solutions and a SAN Architect that focuses on Data Availability Solutions.
In most IT circles, the Master ASE (which hearkens back to its predecessor program at Compaq) remains a widely recognized and highly regarded credential in and of itself. Big Blue's other offerings are more tightly focused on specific platforms like DB2 and WebSphere. These specialized certs are designed to appeal to professionals who use those tools, but they are relatively unknown outside those narrow circles.
Broad exposure to server availability and training in the fundamentals are great. Try following up on that with platform-specific training from specialty vendors. Their products will help you focus on planning, designing, implementing, deploying, maintaining and troubleshooting highly available systems and services.
When it comes to getting availability projects designed and off the ground, your team may not have had the time to become sufficiently trained to take on the projects. Consequently, you may be forced to deal with the expensive prospect of hiring a consultant. If you opt for the consulting route, be sure to take advantage of the consultant's expertise. As full-time professionals in the high-availability field, they are a great resource for developing an ongoing staff training and development program.
This was first published in May 2005