• 05/02/2014
    9:00 AM
    Lee Badman
  • Lee Badman
  • Commentary
  • Connect Directly
  • Rating: 
    0 votes
    Vote up!
    Vote down!

Cloud-Managed WLAN Is Hot, But Not For Everyone

The market for cloud WiFi has grown exponentially, providing a range of options suitable for small organizations, but not so much for large enterprises.

The cloud-managed WLAN market continues to expand with established vendors and startups rolling out a variety of services. But despite the growing number of options, cloud-managed WLAN isn't practical for everyone. The market has yet to offer the types of options that make it feasible for large-scale enterprise WLANs.

Cloud-managed WiFi is hot. Recent entrants into the market include Ruckus Wireless and the newcomer Relay2. Though Ruckus's service is an order of magnitude less expensive and more feature rich than Relay2's, the fact that new players are showing up in a crowded field shows that we're not yet at vendor saturation.

The cloud-managed WLAN market has evolved rapidly from just a few years ago, when Meraki and Aerohive stood in defiance of all the controller and fat-AP enterprise WiFi players. In 2011, the niche provider PowerCloud's launch was significant in that it was arguably the first real stab taken at marketing cloud-based hotspots directly to retail and hospitality sectors. Though this seemed a bit strange back then, PowerCloud actually blazed a trail of sorts with the promise of easy-to-install, low-cost, manage-from-anywhere WiFi for chain stores, restaurants, and hotels.

Today, we have everyone from Aruba to Xirrus with a cloud WiFi story, though seldom are any two services exactly alike. Meraki (now part of Cisco) and Aerohive Networks offer end-to-end cloud-managed IT environments with tightly coupled WLAN components, while other vendors offer simple cloud-managed WiFi access, or WiFi plus some feature set, with no switch or security appliance options.

Small, established environments are usually updated to cloud-managed WLAN models with relative ease by virtue of their small scales. I have upgraded a dozen small branch locations to cloud-based WiFi, and I have become both a believer and an advocate for the paradigm. These spaces were changed fairly easily because they were small, with no AP more than a single switch hop away from its gateway.

But I also have a 4,000-AP-strong enterprise that can't easily move to the cloud. It's not for lack of want, but because of something that's missing in the WLAN industry's overall approach to cloud-managed WiFi architectures.

Today's cloud wireless systems feature high-performance 802.11n and 802.11ac access points, just like I have in my enterprise WLAN. Cloud systems can provide all the modern security features I need, and they provide a range of analytics and reporting that come close enough to my on-premises system. But here's the problem: I can't simply migrate my current-generation, cutting-edge, multimillion-dollar WLAN to the cloud for management. I'd have to rip and replace the whole enchilada.

I'd have to abandon my controller investment and replumb my network topology with VLANs and such, because my "thin" Cisco APs encapsulate VLANs in a CAPWAP tunnel, while cloud-managed APs emulate legacy "fat APs" to which you run trunks. Note that this isn't a problem exclusive to Cisco deployments.

The missing piece for customers like me? There's no "hybrid" model that lets you keep everything you have in place for controllers and APs but move network management and monitoring to the cloud. This is when many cloud vendors might say, "Yeah, but with our cloud solution, you don't need controllers" -- and they're right. But consider the tradeoff.

To configure the underlying LAN topology to support my 4,000 APs and the number of SSIDs and networks used in my complicated WiFi environment, I'd need to perform hundreds of thousands of VLAN-related switch configuration steps and do things I'd really rather not do with Layer 2. Or I'd need to completely redesign my network to be more L3-centric. In this regard, I'll keep my controllers for now, thank you. Yet I'd love to ditch all my local management servers, and if I had multiple campuses, I'd love to manage them all from the cloud.

Small environments of all types should give cloud WLAN serious consideration, and I believe the cloud is where the future of WiFi generally lies. However, we also need enterprise WLAN vendors to step up and fuse what's great about the cloud with the advantages of controller-based WLANs. Figure out how to manage them from the cloud, as well, with very little effort by the customer, and the cloud-managed WLAN picture becomes a lot more complete.


The cloud WLAN gap

Since this market isn't exactly new, it's surprising that vendors haven't stepped up with technologies that are more suitable for enterprises. I wonder why. It seems like a big market opportunity.

Re: The cloud WLAN gap
Hi Marcia, I see a couple of hurdles, but I don't think they are insurmountable. Big controller-based WLAN with local NMS is a giant SNMP framework. Some sort of local collection with feed to/from cloud would be needed. Also, local WLAN NMS can be bloated, and not exactly peppy built on flash and Java. A general move to HTML5 would have to happen. I'm sure there are other thorny considerations, but the WLAN industry is full of absolutely brilliant minds that could solve all of this if they chose to.
Re: The cloud WLAN gap

There are clear advantages to cloud-managed networking. Sometimes those advantages become disadvantages with age. I've run into this in my home lab. 

I believe most cloud-managed infrastructures have what could be called an offline mode, so that loss of external connectivity will not shut down your network or wireless LANs. Anything that depends on realtime Internet/cloud access needs to be fixed or replaced. 

However, hardware that becomes a virtual brick without the cloud service is suboptimal. Hardware that can't be un-virtual-bricked at all is bad for the environment and the market. I won't name names on either of these but they both are in my home lab at the moment as paperweights. 

I understand the desire to keep revenue streams going, and not have to provide lifetime support to people who buy on eBay, but if you can't replace hardware on your own or if you can't tell if a piece of hardware is a brick or a usable piece of gear when you buy it, it's going to hurt you if you're a SOHO/ROBO/POHO user. 

At least make it possible for your prospective new customers to unlock a piece of gear at some cost. Or offer a trade-in plan for gear you want to remit to the landfill.

Re: The cloud WLAN gap

That sounds really frustrating Gallifreyan. Are there any providers doing the right thing by offering a trade-in plan, or making it possible for cusotmers to unlock the device?

IDC market estimate

IDC last month released some big growth estimates for the cloud-managed WiFi market. According to IDC, as the traditional enterprise WLAN market starts to flatten, worldwide cloud-managed infrastructure and managed services revenue will reach $653 million this year and $2.5 billion by 2018. No wonder so many companies are jumping into this space.

No trunks required

but at vendors like Aerohive you can have also your existing central breakout design. Therefor they are offering there tunnel endpoint solutions which can realize also GREs.

With that you get centralized management out of the cloud (including the tunnel endpoint),a centralized data plane, but a distributed controle plane.

I think you talk about Ruckus SAMS, which is from my point of view only the missing extension for there portfolio. Ruckus tells everybody that they are the best solution for Hotspots and ISPs, but they never offered a good guest management system for such requirement.
Now they go the same way like AirTight goes with there Social stuff.




Re: No trunks required
Hi David, I'm not quite following you. My main point is not one vendor versus another for those shopping, it's that it would be wonderful to not have to shop and simply have controller-based systems, like my own multi-million dollar investment, become cloud-managed. Period. I know the market offers several cloud variants if I wanted to migrate to a new platform. But I don't- I want my existing platform to be updated with cloud-management and for my vendor to release me and my fellow customs from fat, maintenance heavy NMS while preserving the WLAN systems we have.
Re: No trunks required

Hi Lee,

sorry I misunderstood your main point.

I only wanted to make a comment regarding the central breakout and existing switch infrastructure.


Cloud Wifi Revisited

Interesting conversation. In my research, I am seeing really only 3 vendors with true cloud architecture today: Mojo Networks, Aerohive and Meraki. The rest have strap on cloud architectures where you can really see disadvantages. I am interested to see larger Enterprises make the shift to cloud managed WLAN - probably more of a culture change (from archaic Cisco controllers) vs. a real technical issue. For the record, I am very impressed with the innovation from Mojo Networks and the solution their are gearing for Enterprise environments...