ADI More Bad Advice: Network Switches

By Brian Karas, Published Mar 04, 2016, 12:00am EST

Run with scissors?  Play in traffic?  Trust ADI for networking advice?

The last we checked with ADI, they were promoting bogus numbers from a shady research firm.

Now, ADI has turned back to the technical side, with bad advice on network switch selection. In this post, we examine what they recommended, what is wrong and 5 reasons why you should consider the alternative.

ADI ******

*** *********** *** ***** ********* ******** *** IP ****** ************.  ***** logic ** **** ******* switches *** *** '************, *********, ***** ******* and ** ***** “**********-*****” networks', *** **** ********* switches "**** ******* ***-*** **** a ***** ** ********* intended ** **** **** with **** ****** ** devices ********* ** *******." 

******, *** ***** ************, in *** ****, *** recommends ******** ** **** 3 ****** ******** - ****** of *****, ****** ****** and ****** *****.

IPVM ******** ** ******** / ********

*** ********** *** ***********, see: 

5 Benefits ** ******* / ***** ********

***** / ******* ******** *** make *************** **** ****** when ******, ****** ****, money *** *********. *** this ** ********** ********* in ** ***** ******* where *** ****** ********* load, *** ******* *** potential ************ ****** *** be * *******.

***** ** ** ******** set ** ******** **** a ****** ** ******** to **** ** ***** to ** ****** * "managed ******", *** ***** is ** ********** ***** of ******** ***** ******** to ** "***** ********" that ********* **** **** features **** * **** managed ******.

**** ******* * ******* defined ******* ***, ***** are * *** ****** that *** **** **** on ****** ***** ***** or ******* ****** **** can ** ********** ** security ******* ************ *** troubleshooting:

*** *********

**** *********, **** ***** **** to ****** *** *** interface ** *** ****** gives *** ** ********** that ** ** ** and ******* *** *** outage ** *** *** to *** ****** ****** being ****.

**** ******** / ****** Monitor

*** *** *** ** a ****** ***** ***** are ** ** ****, and **** ****** **** remember **** ** ******* in ***** (******** **** you *** **** ** properly). 


*** ********** / ***************

* *****/******* *** ****** should, ** * *******, show ***** *** ***-**** power ***********.  **** ****** will **** *** *** power ***** *****, ****** it **** ** ******** reset * ******, *********** saving *** * ***** roll. ********, ** ******* are ******* *** **** power, * ****** *******, this *** **** * tech ****** **. ******* this, *** **** **** have ** *****.

**** ******/*******

******** **** **** ******* between *** ****(*) ******** in *** ******* ****, a ******* **** ****** you ** ***** ** the ******* **/**** * device **** * **** like ********* **** *** ** **** helpful **** ***'** ****** to ****** *** *** a ****** ***'* ******* to * ***, ** why *** ******** ***** alert ** ******* ** send

**** ********

**** ***** ** * few *********, **** ******** **** prevent * ***/******* ****** from ***** ******* ** the *** ******* ******** it ** *** ****** interface.  **** *** **** prevent ***** ******* **** being ******* **** *** network

******** ********

***** *** **** ******** of *****/******* ******** **** probably ***** **** ** used ** ****** ** more ******* ********:

  • *** ******** - ** the ****** ** ********* to *** ****** ******* it's ******** *** **** need ** ** ******** with ******* ** ******* settings
  • **** ******* - *****, on * ********* ****** we ****** **** ** get ** ** ***-**** tagging, *** ** *** be ****** ** *** want ** *** * camera ******* ** *** same ****** ** ******* traffic
  • **** ******* - ******* up * ***** *******/********** tool **** ****** ** simple, *** ** ***** be ******* ** * small ******, ** **** cases **** ******* **** go ******

Modest ***** *******

***** *** ******** ** smart *** ********* ********, *** ADI's ********, *** ***** ***** that ***** ** * huge ***** ******* *** a ******* ******.  **** years *** *** ***** delta ******* ******* *** unmanaged ******** ***** ** ***** significant *** ***** ** is ****** ******* ********** compared ** *** **** for ***** *************** ** video ******.

**** * **** ** some examples ** ******:

  • ******* ***** *** - **** *** * $20 ***** **********.  **** switch *****'* ** ***-**** PoE ***** ******, *** it **** **** * basic *** *********, **** statistics, **** *********, *** some ***** *********** *** detecting *** *********** ***** faults.
  • **** ** ** * Port (*** ***)**-**** ***** ********* $***, *** * comparison *** **, ****** with **** * *** ports: **-****-**-*******, **** ** **** of * ****** ********** and ** ***** **** a $** ***** **********.  The ******* ****** ***** pay *** ****** *** first **** ** ******* for ****** *********** *** prevented * ***** ****, or ** ** ***** ~********* ** ******** ****.

Deciding ** ******* ** *********

**** ******-********* ******* *** likely ****** *** *** of * ******* ******* switch.  ***** *** ***** benefits ** ******* ********, and *** **** **** they *** ******* ********* in * **** ******** indicate **** *** *** intended **** *** "***** ** ********".

***** ** ********* ****** is *** '*****' *** integrators ****** ********* ***** the **** ******** ** these ******** ******** ** their ****** ***** *******. Indeed, *** ***** **** pay *** * ****** more *** * ******* switch.

Comments (37)

"Run with scissors? Play in traffic?" Haha, learning the hook lines for articles early. :)

But yes, having a minimum of a "smart" switch can save a lot of time and trouble for everyone all the way around, unless your clients are right next door and you get to bill those high hourly onsite rates.

I never run with scissors; mainly because I never run.

Box shifting companies ADI should stay out of IP security systems. In the world of IP video / access control / fire systems, more than basic parameters are required to be understood than plugging your IP / reader camera into an un-managed switch, well that is if you want it to perform as designed. Managed switch specifications such as Switching Fabric, Packet buffers, DRAM, Transfer rates etc are all important factors when doing your network design and calculations for traffic flows etc. Having the switch support industry standard protocols such as VLAN tagging, Flow control CoS etc. Lastly, if your units will be in a harsh environment, the environmental specs will be important, don't put a regular Netgear unit out in a box on your perimeter where heat and rain will pound it throughout the year, go industrial.

My two cents, or maybe it was five .....

The dilemma, commoditized product with margins that don't cover "true" technical support.

This industry has been "bastardized" like so many others, leaving everyone chasing diminishing returns, in turn cutting overhead while in survival mode.

You get what you pay for.

Up front savings, long term headaches and backend $$$ to clean up and correct deficiencies.

"commoditized product with margins that don't cover "true" technical support."

Agree generally but, in this case, would not ADI make more money selling smart / managed switches rather than managed ones? And reduce support calls?

My personal theory is that the average person buying networking gear from ADI isn't super tech savvy. This theoretical customer has heard about managed switches, and probably about how IP camera networks need some kind of 'special' gear.

ADI is promoting unmanaged switches primarily because:

1) They don't want their customers being scared off by IP systems

2) They don't want to deal with tech support calls from people messing with QoS and VLAN settings, or possibly twiddling a bunch of settings and causing problems, returned equipment, etc.

ADI's recommendations here seem clear, the extra $5 they'd make on selling a managed switch would cost $20 in tech support.

Agreed. A large portion of the ADI customer base consists of the traditional "alarm guys". They largely are looking for "plug-n-play" systems that get them in and out. They aren't doing complex integrations. Even if they purchased a managed switch or router, they would just plug it in without configuring it anyway. Hell, lucky if they even change the default password on the cameras. Of course these guys are creating countless thousands of security vulnerabilities. Where all this leads remains to be seen.


As a manufacturer I would say ADI doesn't worry about support calls as they will tell their customers to call the manufacturer for support.

"As a manufacturer I would say ADI doesn't worry about support calls as they will tell their customers to call the manufacturer for support."

You win.... :)


I used to work for a small VMS company, and I found myself wishing that all distributors did exactly that. Every time a distributor did tech support, it added what ever time they spent on the call and multiplied it by three to the time needed to fix it. And every call got escalated to us eventually.

And because it was a distributor, what would have been a simple call to a level one support person, ends up being automatically escalated to a sales engineer. So for the manufacturer side, it's more expensive in terms of both time and personnel involved.

"Every time a distributor did tech support, it added what ever time they spent on the call and multiplied it by three to the time needed to fix it."

4, thanks for the feedback. What did the distributor do to multiply the time needed to fix it?

It varied by incident and distributor. In some cases it was playing with settings randomly and not turning them back after they didn't fix a problem which caused other problems. Randomly installing and uninstalling things which caused things to be installed with the wrong flags or caused missing dependencies. Those made up the bulk of them. Just the kind of things that tend to make diagnosing the original problem harder, and then you have to be very through about checking for other stuff. So it ends up being time consuming and something a level 1 tech could have handled in the first place. It's not that their techs were particularly stupid, although the changing stuff and not setting it back speaks to a training issue, it's just a situation where a little bit of knowledge can be dangerous.

There was the occasionally really dumb ass thing like the time they told us an upgrade broke everything when the upgrade in question was a different VMS. But stuff like that is was more the outlier then the norm.

4, thanks, great feedback, though kind of sad...

Hiring personnel with network backgrounds sitting on the phone walking(instructing/teaching) them SMB through the setup doesn't take long before they've eaten up the margins...

I owned a nationwide CISCO integrator and this is the business CISCO wanted the channel to handle, because it wasn't profitable for them...well it's not profitable for most unless they can charge and this industry doesn't appear to understand how to charge for the "intangibles" has they fight off the incursions from China with HD Analog, etc..

"incursions from China with HD Analog, etc.."

Speaking of HD analog, if ADI wants a dumb system, why not just go HD analog? Seems to be less trouble for the average ADI dealer than connecting a networked system with a dumb switch.

I'll add two more capabilities often found on managed switches as other things that you might want or need to have, although like QoS, typically only come in to play in more sophisticated networks.

  1. STP - Spanning tree protocol to prevent broadcast looping.
  2. IGMP Snooping - For multicast.

One thing I sometimes like about dumb switches though, is their dumbness! When troubleshooting a network; you know right away it has 0 configured capabilities, so you if you plug in a patch and you see no lights, either there is no signal or the switch is fried. Unlike the managed switch that you know you will need to log into to check that it's not setup some crazy way.

Though had ADI recommended managed switches for "residential and light commercial" use, would they not have gotten ripped even worse, because of the perception of needless upselling?

Do you think that light commercial users are being 'needlessly upsold' when recommended to use a managed switch?

Depends what you mean by light.

Do you think a managed switch is necessary for an 8 camera system?

It's not necessary for the basic networking. There is practically no need for QoS, VLANs, multicast support or such things in an 8 camera system.

The managed switch is *beneficial* and worth the price delta in an 8 camera system when you need to do any testing or troubleshooting.

It's kind of like insurance. If you pay for it and never ever use it, then you would have been better off without out. But if you use it even once it can be cost-effective in terms of total savings or preventing an unexpected large financial hit.

Not at all. In fact, ADI carries a line of managed network devices (Luxul) specifically marketed toward residential and light commercial applications. Despite my feelings toward ADI themselves, Luxul as a brand is actually quite good. It's almost unfortunate that they appear on the line card of such a terrible distributor.

I think that the kind of person who would turn to ADI for networking advice is exactly the kind of person for whom managed switches would be more trouble than they're worth.

In a way, this article is kind of a test; if you're uneducated enough to believe it, you're uneducated enough that it constitutes good advice.

Many years ago the price delta between managed and unmanaged switches could be quite significant...

That TP-Link switch is definitely a great value, I think I would prefer paying a little more for that over a dumb switch. Thanks for listing it!

To be realistic though, the comparison should be made from the TP Smart switch to the to SF1080p ($44) instead of the SG1080p ($72). It's true that the SG is GigE, like the smart switch, but when doing small installs without network aggregation, 100 Mbit should be enough.

As for the 4-port vs 8-port POE, that is definitely a plus for the smart switch, but the total POE power budget is the same 53W for both switches...

It can also be said that you get what you pay for. While TP-Link has a great price point (and products that aren't terrible in many respects), they aren't applicable in every situation. They are highly insecure and have frequent security vulnerabilities (a problem since most installers will not perform firmware updates with installation). They are also disallowed in some installs since the company is Chinese-owned and subject to the same regular issues as most other Chinese brands in the space. (backdoors, CSRF vulnerabilities, etc.)

but the total POE power budget is the same 53W for both switches...

I saw that too. But there is no shortage of cameras that use 5w or less, so 53W over 8 ports still gives you plenty of headroom.

I agree, it's better to be able to spread the power out among as many ports as possible, though there are some power-trippers who would make the argument that if you don't have full power on every POE port, you will regret it.

Here's an untouched angle so far, of what type do you consider the built in POE switches in the now ubiquitous Hikua NVR's taking over the world to be?

Smart or dumb?

It’s best to recommend an unmanaged plug and play switch that can be installed quickly with no programming required.

ADI continues with this. I kid you not, ADI is recommending their dealers to become IT consultants.

Possibly in their defense, they are telling dealers to "look outside the box" and see if you can grab something low hanging.

On the other hand.....what a potential nightmare and liability!

This reminds me of business owners with no IT knowledge hiring an IT manager. Just sound like you know a lot and if they bite, reel it in. Until you get fired and leave a mess behind.

...see if you can grab something low hanging...

like the cable...

In all cases, always look at the network cabling system for functionality and future connectivity needs prior to making a recommendation to your customer.

Customer: I think the cable should be good, it's a high grade UnTwisted Pair, you know UTP, but have a look.

ADI consultant: Uh, oh. That's what I was afraid of... Managed Cable.

Note: ADI reports that their switch post we criticized was their best read one of the year. Thanks IPVM members!

This is an older thread, but I completely agree with Brian's summary of switches. Only $10 per port... I think that support for GigE is in the same category as management and really should be the default if at all possible. GigE provides several benefits that are often overlooked. Real network management setup and configuration like QoS, VLANs, STP, etc... can require more knowledge than most installers will have. Certainly the general field installer. However, now that Rapid Spanning Tree (RSTP), 802.3af/at PoE and 100/1000 Auto-negotiation is so reliable and easy to use, the default settings on many switches are good enough to keep you out of the ditch. I always tune things like that, but it is rarely if ever absolutely required...

•Cable Diagnostics - GigE switches today with almost any management ability are manufactured with the actual chip that touches each port capable of basic cable diagnostics. TDR, resistance, etc... That goes a long way toward trouble shooting the cable installation. 100Mbps switches do not usually offer this feature. They only need 2 of the 4 pairs of wires and cannot test them all. GigE uses all 4 pairs and generally can/will allow testing of the cables remotely.

•GigE bandwidth usually needs no managing in a small environment. Since most cameras are still 100Mbs, there is no way to saturate any link to the NVR which is usually GigE in a small environment. You still need to manage the bandwidth, but if a dark/snowy/rainy outdoor night scene causes several cameras to exceed what you are expecting and all you have is a 100Mbps switch for the NVR, the port can be so congested, that you cannot remote into it to understand the issue.

•IP Cameras are usually the slowest NIC you install. Except for very small printers or other ultra low bandwidth network attached devices, everything else you connect would be GigE. Routers, Access Points, Wireless Bridges (for connecting remote cameras), PCs, NVRs, NAS for archiving.... They would all generally be GigE.

If you REALLY are certain that the switch is never going to have to have anything but those 4 cameras installed and the NVR just lives on the same switch then fine. In a very tiny installation a small 100Mbps switch might be fine, but for long term support, and for a reliable/flexible installation, the few dollars more for GigE is well spent. Entry level web managed GigE PoE switches currently start at $10 per port. Only $10... That is less than the sales tax on each camera... Yes, we generally should know more about the network and correct installations methods. (This is IP Video after all..) Yes, we should understand camera bandwidth and the video settings to manage it, NVR storage parameters and reasonableness, Pixels per foot for images, Lux levels for image brightness clarity, lens choices for coverage, focus and stuff we all know at least a little about. (I still learn things every single day even after years of practice...) However, without all that knowledge or at least until we gain more of it, a simple managed PoE GigE switch can help save your project/customer from failure and give you insight in minutes that otherwise can take you hours. Also, keep a spare that you know works and learn to use it, just like you do UPS units, PoE injectors, cameras, crimp tools, cable wire testers, volt meters, etc... This was probably too long, but wired network bandwidth today is far too cheap to waste time on conserving this way. 10-20 years ago that was not true. Your time is worth much more to that to yourself, your customer and your friends/family. It just isn't cost effective to worry about 99% of the time. (This isn't true for wireless bandwidth, camera bandwidth, etc...) I am not saying to waste money or bandwidth, just spend the effort where it makes a difference.

Managed vs. Unmanaged, 100Mbps vs. GigE...

Those are not places to spend time arguing about at least to me... Spend your time/resources/etc.. where you can make a real difference to the customer. More time to train them, more time to understand their needs, more time engineer the NVR/NAS, more time to make it an attractive and functional system, etc... -scene from your office- Customer calls at 5:30 with a camera that is dropping out. It is your business, so you answered the phone even though it is after hours. You want to help, but they are across town and 5:30 traffic means it is 1-2 hours there and back in traffic. You just webbed into the switch, saw the port with some error counters that are not 0 and it is the camera port in question. So next you ran the cable test. The switch cable test says one pair is shorter than the others by the length of the 10 foot patch cable just installed last week... You asked them to just unplug and plug back in the camera cable that was not clicked in all the way. After the camera reboots it stays online without errors and stops dropping out. Customer is happy, you are happy... Happy emoticons :-) (emoji for you youngsters) all around.. Now go home and play ball with your kids, hug your spouse, have a beer with friends or do something else because you didn't worry about the $10 you had them spend. Just thoughts...

Cory, when you say GigE, are you sure you mean the GigE machine vision and industrial camera protocol?


Or do you mean Gigabit Ethernet?

....when you say GigE, are you sure you mean the GigE machine vision and industrial camera protocol

Are you sure you do? :)


1000BASE-T or Gigabit Ethernet.  It is just shorthand for the longer to write and many 802.3 standards of the ISO layer 1 and 2 for Gigabit Ethernet.  I can be too wordy at times and was trying to keep it shorter...

Thanks for the clarification question. 

GigE is primarily a protocol architecture that defines how a Gigabit Ethernet network connection (or faster) can be used for transmitting much higher fidelity and frame rate video streams for high speed capture that would not be considered H.264/H.265 friendly.

I am not aware of any GigE Vision network switches or really any network infrastructure switching hardware that is GigE Vision specific.  Mostly they just require non-blocking wire rate switching with sufficient buffering and rate management that continuous streams of data can be maintained reliably end to end.  It generally requires switching capabilities along the lines of iSCSI.

Does this clarify things?

Hrrmm, great, more duplicate terms that can mean different things. I didn't know the "GigE" term caught on in the IT world, too. I'd probably prefer GbE to keep it separate, but that's my own preference.

Read this IPVM report for free.

This article is part of IPVM's 6,731 reports, 908 tests and is only available to members. To get a one-time preview of our work, enter your work email to access the full article.

Already a member? Login here | Join now
Loading Related Reports