Many service providers have decided to build multivendor Multiprotocol Label Switching (MPLS) networks, usually with pretty good success rate. MPLS technology and protocols has been standardized for years and the major vendors have worked out most of the interoperability problems. Choosing a small vendor with insignificant market share is obviously a riskier proposition.
However, you should be aware that as soon as you start building multivendor networks, you become your own system integrator. This approach usually works well for large service providers with knowledgeable technical teams who can quickly and efficiently troubleshoot potential interoperability problems; in these cases, the potential acquisition savings far outweigh the increased support costs or lost revenue. The situation might not be so rosy for new entrants or smaller service providers.
When evaluating whether you should build a multivendor network, also keep the following considerations in mind:
- Do I know the vendors? Do I know the quality of their support? Can I expect (with reasonably high probability) that they will be able help me after closing the sale?
- Do I have the expertise to do the integration work myself?
- If I need expert advice on Vendor X, can I get it locally? How many engineers do I know that are fluent in working with Vendor X or certified by that vendor?
- If I need more engineers to support equipment from Vendor X, how easy would it be to find and hire them?
Last but not least, start small. Introduce a few boxes from an untested vendor into your network and see how they work. Test the quality of their support -- after buying them; pre-sales support sometimes promises, and even delivers, heaven-on-earth -- and decide on your long-term multivendor strategy only after gaining significant hands-on experience.
This was first published in June 2010