Long time ago in a podcast far far away Greg and Ethan pondered whether networking solutions need to scale or not, and obviously one cannot disagree with their generic conclusion that enterprises need just-good-enough solutions and not Google-scale architectures.
However, do keep in mind that:
- Not all enterprises are of the same size. Something that works wonderfully for a single-site SMB shop might not work for a larger enterprise;
- Sometimes a solution doesn’t scale even to the small enterprise size (OpenFlow-based load balancing comes to mind) or is so broken that it should never have been allowed to escape from PowerPoint slides;
- A solution that can survive Google-scale deployment is obviously better engineered and implemented than a hodgepodge of kludges thrown together until the mix worked in a PoC lab, so you might be better off using that solution than an “enterprise-grade” product designed to exploit lack of inter-team communication.
In any case, the right answer (as always) is it depends, and if you want to go beyond that (and you should) you have to:
- Understand the requirements (both functional requirements and expected traffic load);
- Understand the architecture and the limitations of the solution(s) you’re considering for your deployment;
- Select the optimal solution that meets your requirements.
BTW, following this process might make you worthy of having the architect or engineer job title. Buying an enterprise product because you trust your beloved $vendor makes you a consumer.
Getting hands dirty
Are you building a new data center and trying to figure out which vendor to go with? You’ll find an overview of potential requirements to consider and fabric architectures in the Data Center Fabrics webinar. Prefer a more interactive approach? You’ll get plenty of that in the Building the Next-Generation Data Center online course.