It may sound cliche, but I firmly believe any network is only as good as the tools used to manage it. This is another one of those areas where the size of the network bears on the importance of the management system in direct proportion, and when the network happens to be wireless the stakes are even higher. Ineffective or poorly designed WLAN management can be a bigger pain than keeping up the actual WLAN itself, and certainly comes into play when lofty concepts like TCO are being considered. And it’s only getting more complicated.
As important as the Network Management System (NMS) can be for many of us, it’s also a curious beast in a number of ways. There doesn’t tend to be a lot of marketing around NMS products, and it’s hard to elevate management systems to the same level of hype that usually accompanies the introduction of new wireless access points or controllers. I’ve yet to see a whitepaper put out that claims “our NMS is 10X faster at cranking out client connectivity reports than the competition!” but these infomercial-style performance rants for WLAN hardware are pretty standard fare. It’s easy to “sell” access points because you can key on speed or number of spatial streams and build a sales pitch around that. But with NMS, the industry hasn’t put a lot of energy into trying to get people fired up about what amounts to an accessory to the WLAN (albeit a critical one.)
My own experience goes back to the days when Aironet APs were typically too few to warrant NMS. Over time I have used and wrote about Cisco’s WLSE , Airwave’s AMP, and other wireless managment systems. Back in the day, I migrated WLSE to WCS when “thin” became “in”, and hoped like hell that Cisco would ultimately buy AirWave (they didn’t, and Aruba is better for it). I had to disqualify otherwise interesting WLAN vendors from RFC’s because they couldn’t deliver effective NMS solutions along with hundreds or thousands of APs, and I’ve watched cloud vendors crack the NMS nut with varying degrees of success to date. It’s a subject that is near and dear to my heart, as I support 20K+ wireless clients in a growing number of locations in the US and overseas. When NMS is done right, it’s easy to appreciate. When it falters or becomes more maintenance-intensive than the WLAN it’s supposed to support, I get seriously frustrated (or downright angry at times).
As we get into ever bigger networks (both single-site and distributed) and an era where Wi-Fi has become the predominant client access method for many, my own reliance on (and expectations of excellence from) WLAN NMS systems continues to grow. The added dimensions of .11ac, location-based analytics, and a variety of portal and security features mean the NMS development folks have their hands full. I don’t envy their challenge of having to put together solid, functional UIs that also stitch together ever more information on systems that are starting to require their own little corner of the data center when you measure them in terms of cores and memory required. Areas that scare the hell out of me going forward include horribly complicated licensing paradigms for the different functional components that make up well rounded NMS, and “unification” of LAN and WLAN network management that puts (for me) hundreds of switches into the same framework that manages thousands of APs. Again, scale matters bigtime in both of these regards.
If I could rub the lamp and get the Wireless NMS Genie to give me some wishes, I’d ask for these:
- Wireless vendors, please remember that NMS can’t be an afterthought. We seriously rely on it.
- NMS developers can’t just slap together a bunch Java and Flash windows and push NMS out the door. NMS has to be responsive, reliable, and scalable for big environments. Developers have to be plugged in to systems experts.
- No WLAN vendor should ever put out APs or controllers that can’t be fully managed by their NMS (the Cisco 5760 example is pretty bad)
- Care and feeding of the NMS can’t take more time and effort than keeping the WLAN healthy. Upgrades to NMS can’t be overly frequent, depend on weird relicensing paradigms, or be so buggy that weeks are spent on the phone with the vendor just to do what should be routine maintenance. Put reliability above constant feature roll-out.
Given that I’ve never had to develop an NMS, I realize it’s easy to armchair quarterback on the topic. But as a paying customer, by golly, I want my money’s worth and have a front-line view into what works and doesn't. Collectively, we’re too far down the wireless road to accept anything less. How about you- do you have wireless management scars or success stories to share? Do you agree with the importance of good wireless NMS?