Wi-Fi Roaming Analysis Part 3 - Measuring Roam Times

This article picks up where we last left off in our discussion on Wi-Fi roaming. In part 1, I covered how connection control occurs in Wi-Fi, the importance of roaming, and what conditions are involved in triggering a client roam. Then in part 2, I dived into the many variations of Wi-Fi roaming and how they each work. Now that we know the background on how Wi-Fi roaming occurs in multiple different scenarios, it's almost time to dig in and get our hands dirty by actually capturing packets and measuring client roaming performance.

But before we an do so, we have one more topic to cover, namely - how to actually measure the roam. This may seem trivial, and really it isn't that difficult of a subject. However, it is important to establish the methodology we will use to provide consistent, repeatable, and comparable results. This will enable us to accurately compare roaming performance between different types of clients as well as across firmware, driver, and configuration updates on the same client or WLAN infrastructure.

Wi-Fi Roaming Analysis Series:
  1. Part 1 - Connection Control and Importance of Roaming Analysis
  2. Part 2 - The Many Variations of Wi-Fi Roaming
  3. Part 3 - Methods of Measuring Roam Times
  4. Part 4 - Analysis with Wireshark and AirPcap
  5. Part 5 - Analysis with Wildpackets Omnipeek (coming)
  6. Part 6 - Tips for Roaming Performance Improvement (coming)
Measuring Roam Times
There are several different methods by which we can actually measure the roaming event. Variations exist because organizations, wireless professionals, and software tools have different views on what constitutes a completed roam (from beginning to end). Different methods of measurement may be applicable in different scenarios, but what is important is to maintain consistency in approach in order to establish a baseline in performance and be able to accurately compare results to one another.

The following are common methods used to calculate the duration required for a client to roam from one AP to another AP:
  1. Between 802.11 Encrypted Data Packets to/from the client on the old and new AP
    This method focuses on the analyzing the impact of roaming delay to the application(s) running on the client device. By analyzing the amount of time between the last data frame transmitted on the "old" AP and the first data frame transmitted on the "new" AP, we can get an idea as to the latency that is experienced by applications. This can be useful to understand how roaming latency might impact software development and internal application timers that may result in timeout errors or otherwise disrupt network applications.

    However, one drawback to this method is that it reflects not only the actual wireless roaming latency, but also any idle time between frame transmissions if the application does not have data buffered for transmission. This can inflate perceived roam times based on application behavior. For example, many applications are bursty in nature, and measuring roam times between data frames could result in a large amount of time being included that is simply client idle time since no frames have been sent by the application for transmission. Even in the "best-case" scenario of where application behavior is consistent, such as a VoIP G.711 call that sends frames every 20ms, this can still cause imprecise measurement. Historically, this has not been a problem when a large percentage of clients did not support fast roaming methods and roaming times were comparatively large. An inaccuracy of 20ms or less would not be a substantial factor in a 500ms - 1sec roam time. But as newer clients support 802.11r and Fast BSS Transition, roam times are likely to be 50-100ms. A measurement error of 20ms could represent 20-40% of the calculated roam time.

  2. Between 802.11 Probe Request through EAPoL-Key (or Association Response)
    This method focuses solely on the wireless roaming latency, removing application behavior from the calculation. The calculation begins with client probing to discover candidate APs to which it can roam, and typically completes with the final EAPoL-Key frame (but may vary based on which type of roam was completed; for instance with 802.11r Fast BSS Transition the final frame required for a successful roam is the Association Response. See Part 2 of this series).

    However, the drawback with this method is that the calculation may be inflated by including the time required for client probing. Since client probing does not always reflect an actual roam event (many clients probe periodically to maintain a list of APs to minimize discovery time when they actually need to roam) and probing behavior varies between manufacturers and even driver versions, this can result in the inability to accurately compare roaming performance between clients or software upgrades.

  3. Between 802.11 Authentication Request through EAPoL-Key (or Association Response)
    This method is nearly identical to the previous method, but it omits client probing from the actual roam time calculation. Many times, when this method is used, the client probing is still listed for informational purposes in order to better understand the client behavior. Therefore, the roam time calculation is limited to the time it took the client to move to the new AP once it decided that a roam was required. Measurement is typically performed from the client 802.11 Authentication Request through the final EAPoL-Key frame (or Association Response, depending on the type of roam. See Part 2 of this series).

    It should also be noted that 802.11 Authentication does not always indicate a roaming event; Wi-Fi clients are allowed to perform 802.11 authentication with multiple APs at once, but can only be "associated" to a single AP at a time. Therefore, look for 802.11 Association Request frames to positively identify a roaming event, then look backwards from that point to identify when the client both probed to discover the AP and performed 802.11 authentication to the AP (different than 802.1X/EAP authentication).
None of these methods are better than the others, simply different ways of measuring the time required for a roam to complete.

It may be helpful for you to reference a simple Wi-Fi connection ladder diagram. At what point do we start measuring roam, and and what point has the roam completed? Which method of measurement is most useful for my analysis?

Measuring Wi-Fi Roam Times

Also, consider which EAP type you have implemented in your network. Different EAP methods require different application flows to complete authentication and can impact the roaming performance for clients that do not support fast roaming. For reference, here is a PEAP authentication packet flow for a client performing a full 802.1X/EAP authentication on a Wi-Fi network.

Personally, I typically measure Wi-Fi roaming times between the 802.11 Authentication Request and the final EAPoL-Key frame (or Association Response). However, this is simply my preference because I often like to compare roaming times between different clients that may be running different applications. By eliminating the Data frames from my calculation, it is easier to directly compare the Wi-Fi driver stacks and radio performance of the clients.

In the next article on this topic, we'll dig into actual packet captures. Stay with me you packet analysis junkies!


The Impact of 802.11ac Gigabit Wi-Fi on Enterprise Networks

802.11ac is being hyped as a dramatic improvement in performance over existing 802.11n equipment. While this will be technically true once all of the capabilities of 802.11ac are available, this is going to take time. Realistically, within the next year, first generation 802.11ac equipment will only offer marginal improvement over 802.11n equipment for enterprise networks. Separating the hype from real-world impact to enterprise WLANs has received little attention.

I recently had the opportunity to provide commentary on the subject for an article written by Lee Badman in Network Computing. Lee did a superb job with the article, as he always does, and provides great insight for corporate WLAN network managers looking for advice.

I'd like to expand a little bit deeper on my viewpoints expressed in his article with this post, and to provide an assessment of the "real-world" impact that most enterprises should be considering.

802.11ac - Wi-Fi Just Keeps Getting Better!
The 802.11ac standard includes complex technology that will eventually allow multi-Gigabit data transfer, but not all aspects of the specification will be available on day one. Similar to 802.11n, which began with two spatial stream devices capable of 300 Mbps and eventually saw maturation to three spatial streams capable of 450 Mbps, 802.11ac will see an initial first wave of products that are capable of 1.3 Gbps with future maturation possibly up to 6.9 Gbps. Whether or not we will actually see 802.11ac products capable of 6.9 Gbps is dependent on hardware enhancements on both the access point and client that are not certain.

First generation 802.11ac products will achieve 1.3 Gbps through the use of three spatial streams, 80 MHz wide channels (double the largest 40 MHz channel width with 802.11n), and use of better hardware components that allow higher levels of modulation and encoding (256-QAM). The 802.11ac amendment also simplifies implementation of standards-based beamforming for manufacturers by focusing on a single form of explicit beamforming, and eliminating the complex number of beamforming methods detailed in 802.11n. This should allow AP and client manufacturers to align on a single interoperable method.

Future releases of 802.11ac will enable even higher bandwidth by allowing up to eight spatial streams, 160 MHz wide channels, and simultaneous transmission to multiple clients by an access point, called Multi-User MIMO (MU-MIMO). In a few years, the realistic benefits for enterprises will likely be four and five spatial stream products with network designs still primarily based around 20 MHz and 40 MHz wide channels. MU-MIMO deserves special attention because it will mark a significant milestone in wireless technology that will allow greater performance through the use of parallel transmissions to two different receivers from the same transmitter. For example, an AP that is capable of 3 spatial streams (3X3:3) could transmit 1 spatial stream to three different clients that are only capable of 1 spatial stream each, concurrently. This will allow enterprises to better-serve large client populations in high-density environments. As described in the Aerohive High-Density Wi-Fi Design and Configuration Guide, the increasing reliance on WLANs as the primary method of network connectivity and for mission-critical services is shifting the focus from designing Wi-Fi networks for coverage to designing them for capacity. MU-MIMO will be a critical enhancement that will allow Wi-Fi networks to scale larger and provide greater capacity to support growing demand.

Real-World Benefits
Most of the current discussion on 802.11ac focuses on large bandwidth improvements that will not be available for several years to come. The short-term improvements are of bigger benefit in small WLAN deployments, such as SMB and consumer homes, where only a single Wi-Fi AP will be able to take advantage of the much wider channels.

First-generation enterprise 802.11ac products that will be available in 2013 represent an incremental evolution of Wi-Fi above current 802.11n products on the market. These first-generation products will outperform existing 11n products, but only marginally, especially in multi-AP enterprise deployments. This is because the majority of bandwidth gain with first generation 802.11ac products will rely on the use of wider channels, up to 80 MHz. Enterprises need to be cognizant of the limitations in spectral capacity when designing and deploying an enterprise WLAN network. Enterprises must be careful to deploy access points on non-overlapping channels, with sufficient signal attenuation between adjacent access points to prevent co-channel interference.

Here is a look at the data rates available with first-generation 802.11ac products capable of 1, 2, and 3 spatial streams. Note that 160 MHz channels are shown only for reference, and will not be available with the first wave of 11ac products.

First-Generation 802.11ac Data Rates
(Note - 160 MHz channels will not be available in first wave of products)
You can find a complete listing of 802.11ac data rates, including 4-8 spatial streams, in the appendix of the Aerohive High-Density Wi-Fi Design and Configuration Guide.

Arguably, the biggest benefit of first generation 802.11ac will be the adoption of 5 GHz bands by mobile devices. This will enable enterprise WLANs to serve mobile devices in greater quantity with better performance due to the mandatory support of 5 GHz frequency bands by all 802.11ac compliant equipment. Today, enterprise WLANs struggle to provide the capacity required to support the large influx of mobile devices like smartphones and tablets. Once mobile device manufacturers begin deploying 802.11ac capable devices, existing 802.11n and new 802.11ac WLAN deployments will be able to provide significantly better services to mobile devices. Due to the enormous popularity and market growth of mobile devices, they now represent a significant portion of the user-base in many enterprise WLAN networks. However, most mobile devices today are limited to the 2.4 GHz band, which is cluttered with interference and offers minimal capacity. This has resulted in under-performing WLANs and an often poor user experience. With the adoption of 802.11ac, mobile devices will be able to take advantage of the cleaner spectrum and the additional capacity available in the 5 GHz bands. In addition, chipset enhancements should allow mobile devices to operate at higher bandwidth and performance levels with better battery life.

The Massive Shift to 5 GHz and the Impact to Enterprise WLANs
802.11ac operates exclusively in the 5 GHz unlicensed bands. In the U.S. there are a total of 25 non-overlapping 20 MHz channels in the 5 GHz bands. While this appears sufficiently large to handle channel re-use in large enterprise deployments, it can be deceiving. When channel size is increased, the number of available channels decreases, constraining channel re-use and making the risk of co-channel interference and WLAN performance degradation higher. Only 10 non-overlapping channels will be available with 40 MHz channel width (similar to the 9 channels available with 802.11n; the difference is due to the addition of channel 144 with 802.11ac), and only 5 channels will be available with 80 MHz channel width. When DFS channels are avoided, which are not supported by a large percentage of client devices and are at higher risk of causing network stability issues, the number of remaining channels dwindles down the only 2. Therefore, it will not be practical for most enterprises to use of 80 MHz wide channels because it will significantly constrain channel re-use that is critical to a high-performance WLAN. This will limit practical performance of first-generation 802.11ac in enterprise environments to 600 Mbps using 40 MHz channels, a far cry from the 1.3 Gbps advertised.

Spectral Capacity versus Channel Width

Enterprises in multi-tenant buildings or in dense urban areas will likely see increased utilization of the 5 GHz spectrum bands, which could cause greater levels of interference and degrade WLAN performance. This is of significant concern if enterprises deploy 802.11ac equipment with 80 MHz wide channels, without recognizing the impact to neighboring businesses.

802.11ac also threatens to accelerate the utilization of 5 GHz spectrum bands by a large majority of enterprises. This could be a double-edged sword, providing the promise of increased performance for individual organizations, while simultaneously congesting the once interference-free 5 GHz bands. This may expose the need for more unlicensed spectrum sooner than anticipated. The timing is impeccable, as the FCC and Congress are currently considering allowing unlicensed use of two additional 5 GHz bands, devising rules for spectrum auctions in 2013-2014 of the 600 MHz TV white spaces, and spectrum-sharing plans in the federal 3550-3650 MHz band that would provide additional unlicensed spectrum for general use. While there may not be an "unlicensed spectrum crunch" today, there very well could be one in the not-to-distant future. More information on the current spectrum policy discussions can be found in my previous blog post about the need for a balanced spectrum policy.

When does it make sense to deploy 802.11ac?
Enterprises that have deployed the latest generation 802.11n equipment pervasively throughout their network can be confident in the investment they have made; first generation 802.11ac only offers incremental benefits over 3 spatial stream 802.11n.

First generation 802.11ac products will be of greater interest to enterprises that are purchasing a new "greenfield" WLAN deployment, growing an existing WLAN deployment with additional APs, or are running on older legacy WLAN equipment. 802.11ac is backwards-compatible with all previous versions of Wi-Fi, so it will be able to supplement existing WLAN deployments seamlessly while providing higher performance and investment protection versus 802.11n equipment. Enterprises that were early adopters of 802.11n may see greater appeal in moving to first generation 802.11ac because their existing 802.11n equipment has already been depreciated over a number of years and they have received their return on investment. In addition, 802.11ac can offer a substantial upgrade in performance to 600 Mbps over two spatial stream 802.11n (300 Mbps), allowing the enterprise to increase performance, capacity, and services offered over the WLAN.

Revolution or Evolution? - Andrew's Take
I'm bullish on 802.11ac. The technology holds a lot of promise to improve enterprise WLAN performance and capacity, especially for mobile devices which are accounting for a larger and larger percentage of our client base with BYOD and Consumerization of IT.

However, we need to temper our short-term expectations. The first wave of 802.11ac equipment will likely not prompt upgrades to existing WLAN deployments unless it is replacing older gear; it just won't provide enough value to justify the replacement of the latest generation of 802.11n equipment. Enterprises will also need to be careful to deploy 802.11ac equipment correctly to avoid harmful interference to their own network and neighboring networks - most notably limiting channel width to 20 MHz in high-density areas and 40 MHz maximum in other common-use areas.

Just as with 802.11n, we will see subsequent releases of 802.11ac that implement additional technology improvements detailed in the standard. This will include 4+ spatial streams (possibly up to 8 eventually), 160 MHz wide channels, and MU-MIMO.

Lastly, we need to ensure the value of unlicensed spectrum is recognized by regulatory agencies throughout the world. The 2.4 GHz unlicensed band is often joked about as being the "junk" band today due to rampant interference and over-crowding. The massive shift to 5 GHz is already underway which will only be accelerated by 802.11ac adoption. We run the risk of 5 GHz over-crowding very soon if more unlicensed spectrum isn't made available.


802.11ac Gigabit Wi-Fi Series:

Cellular MVNO's Increasingly Rely on Wi-Fi Offload to Net Subscribers

Are MVNOs pioneering the next "innovation" wave in cellular? The answer may be 'Yes' based on unique business models that appear to be working.

Cellular MVNOs (Mobile Virtual Network Operators) are gaining attention and subscribers in the U.S. due in-part to cheaper pricing models built around BYOD and Wi-Fi offload:

many new MVNOs are adopting Bring Your Own Device (BYOD) strategies, with SIM-only MVNOs like Simple Mobile, Red Pocket and Ultra on the GSM side, and a new BYO Sprint Device solution for MVNOs like Kajeet, Ting and others on the CDMA network. Sprint’s BYOSD program has the added benefit that no SIM kit or installation is required; the handset is activated simply via its serial number.

With BYOSD, for example, Kajeet offers network-based parental controls, web filtering and location services on recycled handsets. BYOD solves two problems for the MVNO – eliminating handset subsidies and reducing logistics cost (kitting, shipping, warranty repairs and returns). Even where customized handsets are used, MVNOs sell them above cost, eliminating costly subsidies.

Service – airtime and data – costs can also be reduced. With increasing data usage, many MVNOs utilize dual-mode phones (cellular and Wi-Fi) to offload voice and data traffic to Wi-Fi networks, which is increasingly available in homes, offices and businesses. And an added benefit for providers: offloading to Wi-Fi turns off the carrier’s meter.

When Republic Wireless introduced its $19/month unlimited plan as a beta trial, everyone asked how they planned to do it. Republic relies heavily on Wi-Fi networks at home and work, using “hybrid calling” or cellular offload where traffic only rolls to Sprint’s cellular network when Wi-Fi is unavailable. Republic is now shipping a Motorola Android smartphone, running proprietary Republic software (for $259), which completes the no-contract package. And apparently the beta trial worked just fine: the same $19 plan is now available to all.

FreedomPop guarantees 500MB of free 4G mobile broadband data every month, with no data caps or throttling, and attractive plans ($17.99/month for 2GB of data, a cent per MB additional). Customers can earn additional data for each friend referred or unlimited data by engaging in partner promotions. The Freedom Hub Burst, a 4G Wi-Fi router that offloads cellular to wireline and supports up to 10 devices, is free with security deposit. They also offer the Freedom Sleeve Rocket, an iPod Touch case that turns it into an iPhone. Plans include trading bandwidth with other FreedomPop users, and creating bandwidth-sharing communities. Launched on Clearwire, FreedomPop will add Sprint’s LTE network next year.
For those unacquainted with MVNOs, they purchase wholesale access to the larger carrier networks and repackage services with different pricing plans to subscribers.

What I find interesting is the unique business approaches being offered by MVNOs through a combination of cellular data coupled with prioritized Wi-Fi offload when Wi-Fi is available. BYOD is also a new approach in the cellular world, where historically strict control over subscriber client device options helped ensure network performance.

MVNOs are also leading solutions to ease international cellular data roaming.
If you travel overseas and want to maintain your mobile data connection, you’re either going to pay criminal roaming rates or endure tremendous hassles avoiding them. But a new breed of virtual operators like Voiamo are looking to create the first truly international plans.

To put it simply, international data roaming is broken, and no U.S. carrier seems to be lifting a finger to fix it. They seem to prefer the miserable status quo to the headaches required to repair the system. But where the network operators are falling down, mobile virtual network operators (MVNOs) are picking up the slack.

While most MVNOs tend to work in a single country with a single carrier, there’s nothing preventing them from buying capacity on multiple networks in multiple countries and then selling international access to a customer in a single pricing plan.

London-based Voiamo, however, is thinking bigger with a new service called GlobalGig. Instead of just renting you a hotspot and selling you a temporary plan when you travel, it proposes to replace your current country-limited 3G or 4G modem plan with a service that will work in multiple countries with a single pricing plan. Its rates are a comparable to the prices most of the major carriers charge for hotspot plans — starting at $25 for 1 GB a month and up to $50 for 5 GB — but those rates are good for the U.S., the U.K. and Australia.

In the U.S., GlobalGig uses Sprint’s network, while in the U.K. and Australia it uses the 3 and Optus networks, but it is negotiating deals with carriers in other countries and plans to expand its global footprint soon, Voiamo CEO and founder Nigel Bramwell said. Its $120 hotspot — which can connect up to five devices through Wi-Fi — can support networks in 100 different countries, Bramwell said. As GlobalGig adds more carriers to its roster it will periodically send out new SIM cards to its customers, expanding their coverage to new countries.

MVNOs are typically thought of as services that cater to the more price-conscious consumers; those that can't afford more expensive contracts by the larger carriers that lock subscribers into 2-year commitments.

But perhaps it's time to re-assess their offerings and take a serious look at MVNOs for better service offerings than their larger competitors.