Editor's note: In part 3 of our series on rebuilding your critical infrastructure after a disaster, Carrie Higbie...
looks at important factors for rebuilding the original site by addressing the active components on a network.
Assuming that your infrastructure has now been assessed and either replaced or newly installed, the next step is to rebuild the active portion of the network. The active portions would include networking components, servers, storage devices, KVM switches, monitoring and firewall services, etc. This step should be implemented, documented and tested prior to adding desktop services to the network.
Beware of what you can't see
The same factors that can harm your cabling infrastructure can also damage your active components. If the active components were energized at the time that water was introduced into your network, they are probably useless. Be aware that toxic fumes, salt water, clear water and other harmful contaminants may have been around or near your active gear. Humidity alone can render these components useless. Even if they seem to work at first, over time they will likely fail unless they were encased in enclosures that are designed to keep out moisture and contaminants. Even if they were in such an enclosure, it is important to be sure that all openings into the enclosures were properly sealed and that the enclosures were not exposed to the contaminants longer than the warranty for the enclosure allows.
Adding active components to the newly revamped infrastructure may face some hurdles. Today, as companies are keeping active components longer than the two to three year lifecycles that were the previous averages, replacement may mean a complete revamping as those components no longer exist. Also, as the infrastructure was rebuilt, you will want to evaluate the category of copper and fiber components, carrier circuit connectivity and additional services to be sure that you are matching your components. For instance, on the copper side, most companies today are installing at a minimum category 6, 6a, or 7 copper. These are backwards compatible, but as the bandwidth has significantly increased with the newer categories, some components that used to require fiber, may not be services with less expensive copper active components. Some services that used to require single mode fiber may now be available on less expensive multimode fiber components. Components that utilized direct connections with proprietary or legacy cabling may now be connected via standard UTP or S/UTP cabling. In short, work with facilities to assure that the proper infrastructure is in place before selecting active components.
Work with the manufacturer
Firewalls and security devices may be of newer revisions and hardware platforms. Even if the databases and rules were backed up, these may be ineffective with the newer platforms. It is in your best interest to work directly with the manufacturer unless you can be certain that your vendor has maintained their training on the newer platforms. Most manufacturers will either provide direct support or refer you to an authorized reseller with the appropriate skill levels to convert parameters from an older version to the latest release.
The same holds true for routers and switches. There has been much advancement in both of these fields. If you will be replacing a destroyed PBX with the newer VoIP systems, you will want to assure that your active networking gear will support the necessary quality of services. New routing protocols exist that provide better throughput while using less network resources. Whenever possible, it is a good idea to have all vendors for your communications in a rebuild meeting. This assures that all systems will work in harmony when your primary site becomes functional. You may wish to utilize a single vendor solution or multiple vendor solutions. In the latter, getting "Mom and Dad" in the same room can save a lot of time over going back and forth and becoming the middle man in negotiations and project implementations. There are certainly arguments on both sides of the single vendor versus multiple vendor solutions. While a single vendor solution may be significantly more, interoperability issues may be eliminated. Multi-vendor solutions may provide a savings and in many cases will be required to answer all business needs, but will require a bit more coordination prior to implementation.
Ask the experts
Advancements in servers and storage are probably the most significant of them all. Blade servers and storage networks have gotten increasingly faster in smaller footprints and have a wider variety of features even over last year. Alliances such as the BladeSystems Alliance, SNUG (Storage Network Users Group), DMTF (Distributed Management Task Force) and SNIA (Storage Network Industry Association) all have websites with links to member companies and whitepapers on the newer technologies. As the cost of chips is cut in half about every 18 months, you may be able to have a significant boost in functionality for less than your original expenditures on this equipment. The advantage of working through these associations is that they are non-profit organizations that work to assure interoperability and co-existence in their member companies. Many have already tested interoperability. This saves you the hassle of determining what really works together and what tasks are needed for coexistence to occur. Don't forget to take advantage of the wide range of experts on the TechTarget sites. Many of us have already been through similar situations and can share past experiences to assist in your transitions.
Don't forget to look at specialty applications and needs as well such as credit card processing applications, clustering and management applications and operations services. Whether choosing built in management or building on existing management, it is imperative that some means exist to monitor the new products to assure that processes are running smoothly and efficiently. The package that you chose should provide actionable data, dynamic reporting and efficiency statistics. New software and hardware may have different bandwidth and processing requirements than older versions. When your software is updated, intuitive guesses, vendor minimum requirements or assumptions on these needs may render the site useless or ineffective. You will want to monitor all processing for a minimum of the first 30 days on a daily basis with regular health checks occurring monthly for the next several months. Keep in mind that these needs will change yet again when end-users are added to the network.
Coordination has to occur among all teams, vendors, implementers and manufacturers. Due to the vastness of the damage, your time frame for all implementation and testing should take into account the limited resources both internally and externally to your organization. Outside of coordination, testing and reevaluations are key to rebuilding the best services to support your organization as you cut back over to your original site. Part 4 of this series will address adding end users back to your networks to resume regular operations at the primary site.
Carrie Higbie has been involved in the computing and networking industries for nearly 20 years and has taught classes for Novell, Microsoft, and Cisco certifications as well as CAD/CAE, networking and programming on a collegiate level. Carrie currently works as the Network Applications Market Manager with The Siemon Company, where she provides liaison services to assure harmony between active electronics and networking infrastructures. She participates with the IEEE, TIA and other consortiums and works to further educate the end-user community on the importance of a quality infrastructure.