Mesh Without Wires

April 6, 2010

To Mesh or Not To Mesh for Wireless Security and Surveillance

Or Limitations of Point-to-point and Point-to-multipoint Systems for Real-time Video

This is continuation of the discussion on “Why wireless mesh” that got started when I tweeted out my article for SMD last week: Wireless Mesh for Video Surveillance: Why and How. We do get these comments: wireless mesh is expensive, why wouldn’t you use a much simpler and cheaper point-to-point links for video surveillance? Our competitors in the point-to-multipoint space, including fixed WiMAX or ‘WiMAX-like’ systems, echo that by saying that mesh is very complex, multi-hop creates ‘unpredictable environment,’ while point-to-multipoint is deterministic, and therefore better for video.

Disclaimer, before I get underway: low-cost Wi-Fi based point-to-point links have their time and place in video surveillance. If you need to connect a couple of outlying cameras across your parking lot, you can successfully use point-to-point links. We have our own point-to-point product line for exactly these scenarios. However, for the purpose of this article, I’ll focus on deployments with 15 camera locations or greater – a campus, downtown, or industrial site type of environment. The number 15 is somewhat arbitrary, but this is approximately when a lot of wireless providers start having challenges supporting real-time high-resolution video.

PtP and PtMP networks quickly run out of usable spectrum

Compared to point-to-point and point-to-multipoint systems, mesh differentiator is and always has been the ability to “repeat” the signal, and therefore extend the reach of the network beyond the first hop. Eventually PtP and PtMP wireless systems run out of usable spectrum: in order to achieve a clear line-of-site (required at 2.4/4.9/5 GHz), they are forced to radiate from at least one high point (tower or tall building). This fills the airwaves with a chosen band of frequencies, which are limited in number.

Mesh avoids this issue by being able to isolate the RF signal. Radios are typically mounted on street poles where the RF can be directed between buildings and trees in a given pathway back to a final destination, through a series of repeater nodes in a multi-hop configuration. Because the frequencies can be reused along the path, the same limitations of PtP and PtMP solutions do not apply. However, Firetide is the only mesh product that also maintains the low latency, high capacity and distributed architecture required for real-time video, which is why we excel in this space.

Transition to MIMO creates more challenges for PtP and PtMP

It’s also important to note that as the wireless manufacturers move from 802.11a/b/g to 802.11n-based systems, channel size increases from 20 MHz to 40 MHz. That equates to 2 channels in the 2.4GHz range, 1 channel in the 4.9 GHz range, and 11 channels in the 5 GHz range. In other words, 802.11n cuts down the available number of channels by half which will limit the scalability of P2P and P2MP systems even more.

Rigid architecture of PtP and PtMP not the best at handling challenging environments

Urban Canyons Challenging for PtP & PtMP

Urban Canyons Challenging for PtP & PtMP

RF considerations aside, PtP and PtMP system also face other challenges. As I already mentioned, PtP and PtMP systems require tall assets to place the base station in every sector of coverage. We’ve won deals over these providers just because these tall assets (rooftops of privately owned buildings) could not be secured in sufficient quantities. The same challenge applies to industrial facilities, where deploying mesh can be the alternative to constructing towers.

In dense urban environments, even if you do get access to rooftops, the topology of “urban canyons” is such that you cannot achieve LOS from the rooftops to all the camera locations, which requires you to go street level. So that takes us back to street-level coverage with multi-hop, which can only be delivered by a mesh topology. In one city-center deployment – because of dense foliage – we had to shoot under the tree canopy; no point-to-multipoint system can do that.

Of course, you could string point-to-point links and achieve the same mesh-like topology, but then you’ll run into issues of latency and jitter, as you are switching between one PtP link to another. This is most evident in bandwidth intensive applications, such as video, which is why products like Ubiquity or Tranzeo, along with other Wi-Fi products out there, will never be good for more than a handful of cameras in a given geography. The cost advantage is also reduced: you need 4 PtP radios for two links, vs 3 mesh nodes for the same 2 connections. So no dice with PtP either.

Capacity limitations in PtMP lead to high costs of deployments

There are virtually no capacity limitations in PtP space; you can get up to 1 Gig links, from providers such as BridgeWave. So I’ll address the capacity limitations of PtMP systems that we compete with for video surveillance deployments. WiMAX, which is being touted by WiMAX providers as the solution for real-time video surveillance, today delivers only about 30 Mbps of capacity per base station, divided by the number of subscribers. (Theoretical data rate for a fixed WiMAX system is 70 Mbps, so real-world 30 Mbps is generous). A typical standard resolution video camera outputs a 2-3 Mbps video stream, a 1080p HD cameras requires 5-7 Mbps of bandwidth (using H.264), and a 12 megapixel camera can hit as high as 30 Mbps. So you do the math.

According to IMS Research, by 2013 revenue from HD and megapixel cameras will equal that from standard definition cameras (see Day 1 ISC West 2010 Impressions). These systems will be challenged even more as user adoption of HD and megapixel grows.

Mesh is expensive, but vastly less so compared to fiber

To the final argument: “mesh is expensive.” Our solution is not cheap, especially compared to low-cost PtP. But for these types of deployments that I was talking about, low-cost PtP is not the true competitor. The alternative (or the starting point) in the customer’s mind is often fiber – and fiber is vastly more expensive than any type of wireless. In fact, 90% saving off fiber infrastructure is the number I heard just recently at ISC West from our integrator who priced out fiber vs Firetide’s MIMO mesh infrastructure.

In summary, yes, point-to-point links are sufficient for small scale deployments, and you may not need mesh in this case. However, as you move beyond a few links, multiple point-to-point links become very complex to manage, and you’ll ‘gain’ in network management headaches and poor performance what you saved on cost.

For additional discussion on the topic of wireless mesh for video surveillance, see:

Image via Flikr

By: Ksenia Coffman – Connect with me on Twitter or LinkedIn.

9 Comments »

  1. Hi Ksenia, I like this post. I think it fairly explains specific areas of use and disclaims what areas where alternatively technologies may be superior.

    Comment by John Honovich — April 6, 2010 @ 5:08 pm | Reply

  2. Ksenia, these 2 posts have been great and forwarded on to our security salespeople, and will become a standard document for new sales team members so I don’t have to explain it for the umpteenth time

    Comment by Sean Patton — April 7, 2010 @ 8:59 am | Reply

  3. In you claim about point to multipoint being limited bandwidth (e.g. in you WiMax example, which is acurate) please note that just as Firetide is now using the 802.11n chip sets so are many of the PtMP manufactures. On systems we have deployed and tested we have received over 200+Mbps usable bandwidth with radio data rates of 300Mbps. These systems too are using MIMO 2×2 and 3×3 and actual do quite well with NLOS. You are correct though in the fact that most PtMP needs LOS or at least performs best with LOS. The one advantage of mesh has been in municipal applications and the ability to mount on existing infrastructure. Plus the ability to “mesh” around buildings and other obstructions.
    I believe both mesh and PtMP are ideal solutions working together.

    Please keep up with the good posts on mesh! They have been very insightful and have helped non-wireless individuals understand the benefits of mesh as a wireless technology (especially in the video space).

    Comment by Joe Wargo — April 7, 2010 @ 8:49 pm | Reply

  4. This post of yours explain the use of mesh over fibre which will be the next in technology for projects in india.

    Comment by Deepak bhutani — April 7, 2010 @ 11:27 pm | Reply

  5. It was a very good article. However, you are missing the concept that mesh is really PTMP to PTMP. There is no difference between that and using multiple PTP radios to duplicate the same thing. In addition, most mesh radios can really only make one good connection in and out such as in the case of light poles on a street. The radio can see 2 other radios in the chain. Regardless of whether mesh is used or something as simple as a pair of PTP radios, there really isn’t any difference. Now throw in the fact that PTP radios can use multiple polarities and more directional antennas, this might provide a higher modulated throughput over mesh radios using omni-directional antennas at longer distance. If you use directional antennas on the mesh radios, then you defeat the mesh concept completely. There are radios that do this now but I argue that paying $3000-$7000 dollars versus $500 dollars for the same performance is rather difficult.

    Comment by Rory Conaway — April 10, 2010 @ 11:53 am | Reply

    • Good point, Rory, and probably applies to many providers of “Wi-Fi mesh”. However, Firetide’s mesh is much more than Wi-Fi radios arranged via PtP links in a mesh topology. That is why we call it “infrastructure mesh” – L2 distributed wireless Ethernet switch architecture with L3 routing protocol within the mesh. I plan to cover these distinctions in the near future.

      I also do not see using directional antennas as defeating the purpose of mesh. In fact using directional antennas is often the best practice for video mesh – to direct the signal to where it should go, rather than radiating in all directions via omni-directional antennas. You can still build in redundancy by having a mesh node ‘see’ 2 or more adjacent nodes.

      Comment by kseniacoffman — April 11, 2010 @ 3:06 pm | Reply

  6. Actually, I wasn’t referring to other providers of mesh, I was referring to the concept of using all L2 with WDS. As for seeing more than 1 access point per radio, maybe you can explain in a practical deployment how that is advantagous. For example, the directional antenna assumes that the next radio is typical inline but further down in a residential environment. Asusme that the radios are 1/2 mile apart which would by typical for a public safety deployment. Also assume that all AP’s are triple radios configurations meaning the 2 radios are for backhaul and 1 is the local hot-spot AP. Doesn’t matter in frequency. The backhaul can be anything from 4.9-5.8GHz assuming you are in a band with 36dBm output or better to allow for maximum modulation at 1/2 mile. In that environment, the only 2nd AP that you can hit is the unit 1 mile away. Unless there is a radio failure, which is one of the marketed purposes of mesh, there is no reason for hitting 2 radios. However, in my experience, radio failure is very very low, not justifying the additional design burden for trying to compensate every location for scenario. That does have value if it takes more than 4 hours to replace an AP because you need tower climbers, lifts, or cranes to reach one and their availability is unknown. In that environment, there is also no purpose to L3 routing. In an environment were you have a higher density of units and they are closer together, then there is some value assuming several units have LOS connectivity. You are then back to omni-directional antennas which lower modulation rates at longer distances.

    L3 routing has application if you are using multiple SSID’s and you are using a VLAN infrastructure with multiple egrees points to the mesh. I understand that. However, assuming you are only supporting WiFi or Public Safety, a single SSID environment really doens’t care since there is usually only one Ingress/Egress point for the area. Based on the design I described, mesh is irrelevant. It’s a fixed environment.

    Comment by Rory Conaway — April 11, 2010 @ 7:19 pm | Reply

    • Rory, we do not use WDS to communicate between nodes. Our intra-mesh communications are done via proprietary flow-based routing protocol that Firetide designed and patented.

      Comment by kseniacoffman — May 26, 2010 @ 10:22 am | Reply

  7. Skype has opened its website-centered buyer
    beta on the entire world, following launching it broadly in the United states and U.K.
    previous this four weeks. Skype for Web also now works with
    Linux and Chromebook for immediate text messaging connection (no
    video and voice however, individuals demand a connect-in installment).

    The increase of your beta provides assist for a longer set of dialects to aid strengthen that international user friendliness

    Comment by Adriana — July 11, 2017 @ 9:53 am | Reply


RSS feed for comments on this post. TrackBack URI

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

Blog at WordPress.com.

%d bloggers like this: