Re: [WIRELESS-LAN] EIGRP equal cost load balancing over wireless bridges

[Mark Duling]

Thanks David and Eric. It was per destination. Good thoughts I'll look into
before trying it again. Thanks much.

Mark - Biola IT Network Operations

On Wed, Feb 28, 2018 at 7:22 AM, Kenny, Eric  wrote:

> Hi Mark,
>
> Have you tried both links individually?  If they are different vendors,
> one might be handling MTU differently, which might upset your CAPWAP
> tunnels.  Just a thought.
>
> Thanks,
> ---
> Eric Kenny
> Network Architect
> Harvard University IT
> ---
>
> > On Feb 27, 2018, at 7:45 PM, Mark Duling  wrote:
> >
> > Hello all,
> >
> > Is anyone running two pairs of inexpensive wireless bridges–say ubiquity
> AirFiber–and using EIGRP equal cost load balancing over them? It seems to
> me that should be an inexpensive way to support reasonably high bandwidth
> building over redundant links.
> >
> > I ask because a while back as a test I changed a remote building with
> two pairs of wireless transparent bridges (one acting as primary / other
> acting as backup, each pair a different vendor but both wired interfaces
> 100 Mb) from routing over the primary to use equal cost load balancing over
> both. After that we'd occasionally see our Cisco capwap APs disassociate
> and re-associate from our WLCs on campus where they weren't doing that when
> data was routed over a single link.
> >
> > Does anyone know why that might happen with ECLB? Or in any case, is
> anyone successfully using a dual link wireless bridge setup with both links
> active? Thanks
> >
> > Mark - Biola IT Operations
> > ** Participation and subscription information for this EDUCAUSE
> Constituent Group discussion list can be found at http://www.educause.edu/
> discuss.
> >
>
>
> **
> Participation and subscription information for this EDUCAUSE Constituent
> Group discussion list can be found at http://www.educause.edu/discuss.
>
>

**
Participation and subscription information for this EDUCAUSE Constituent Group 
discussion list can be found at http://www.educause.edu/discuss.

RE: [WIRELESS-LAN] Air Time Fairness

[Jason Cook]

Thanks Sam,

Thankfully it’s primarily pre-wave 2 locations where we see this.. Having said 
that yesterday a single client seemed to be chewing up around 50% airtime 
updating steam apps, while the other 9 users were having a bad time…. That was 
wave 2. There are other ways though if needed

--
Jason Cook
Information Technology and Digital Services
The University of Adelaide, AUSTRALIA 5005
Ph: +61 8 8313 4800

CRICOS Provider Number 00123M
---
This email message is intended only for the addressee(s) and contains 
information which may be confidential and/or copyright.  If you are not the 
intended recipient please do not read, save, forward, disclose, or copy the 
contents of this email. If this email has been sent to you in error, please 
notify the sender by reply email and delete this email and any copies or links 
to this email completely and immediately from your system.  No representation 
is made that this email is free of viruses.  Virus scanning is recommended and 
is the responsibility of the recipient.


From: The EDUCAUSE Wireless Issues Constituent Group Listserv 
[mailto:WIRELESS-LAN@LISTSERV.EDUCAUSE.EDU] On Behalf Of Samuel Clements
Sent: Thursday, 1 March 2018 12:08 AM
To: WIRELESS-LAN@LISTSERV.EDUCAUSE.EDU
Subject: Re: [WIRELESS-LAN] Air Time Fairness

One gotcha is the following:

Table 6 Key Features Not Supported in Cisco Aironet 1800i, 1810 OEAP, 1810W, 
1815, 1830, 1850, 2800, and 3800 Series APs
Cisco Air Time Fairness (ATF)

Be warned that, if you're using the wave 2 platforms, ATF is not supported 
according to the latest WLC version 8.6 release notes at:
https://www.cisco.com/c/en/us/td/docs/wireless/controller/release/notes/crn86.html

  -Sam


On Wed, Feb 28, 2018 at 1:15 AM, Jason Cook 
mailto:jason.c...@adelaide.edu.au>> wrote:
We are Cisco shop on 8.2.164 and would potentially move to 8.5 July. (8510 in 
HA, peak 15k clients)

Open to thoughts from other vendors as well 😊

Does anyone know of any gotya’s when enabling ATF with Client Fair Sharing? 
Capacity is one that comes to mind with AVC having clearly caused plenty of 
issues under load.
We are starting on our dev controller 5508’s in HA, so can play with it easily 
enough. But AVC in dev was fine also, hard to test capacity there since 
creating the load is a challenge.


--
Jason Cook
Information Technology and Digital Services
The University of Adelaide, AUSTRALIA 5005
Ph: +61 8 8313 4800
e-mail: 
jason.c...@adelaide.edu.au>

CRICOS Provider Number 00123M
---
This email message is intended only for the addressee(s) and contains 
information which may be confidential and/or copyright.  If you are not the 
intended recipient please do not read, save, forward, disclose, or copy the 
contents of this email. If this email has been sent to you in error, please 
notify the sender by reply email and delete this email and any copies or links 
to this email completely and immediately from your system.  No representation 
is made that this email is free of viruses.  Virus scanning is recommended and 
is the responsibility of the recipient.

** Participation and subscription information for this EDUCAUSE 
Constituent Group discussion list can be found at 
http://www.educause.edu/discuss.

** Participation and subscription information for this EDUCAUSE 
Constituent Group discussion list can be found at 
http://www.educause.edu/discuss.

**
Participation and subscription information for this EDUCAUSE Constituent Group 
discussion list can be found at http://www.educause.edu/discuss.

Wireless RFP - Preparing to Start the Process

[Mike Beane]

Good afternoon,

We've reached the point where we will be starting the groundwork for
putting out an RFP to engage vendors with before next summer.  For my short
time here, this will be the first time I've done this particular system and
even though we've overhauled our virtual infrastructure and Internet
connectivity in the past three years, those were fairly "behind the scenes"
projects.  I believe that this will be our (or at least Infrastructure's)
most visible project to date.  We have both the daily routine of
students\faculty\staff, but also a quarter of our students are residents on
campus throughout the academic year.

I'm looking for anyone who might be willing to share their wireless RFP,
and if you've done this in the last three years, what system did you go
with?  On list or off is fine and either is appreciated.

Thank you,
Mike

*Mike Beane*
IT Infrastructure Manager
*Ph: *207-941-7613
*Husson University*
1 College Circle
Bangor ME 04401

**
Participation and subscription information for this EDUCAUSE Constituent Group 
discussion list can be found at http://www.educause.edu/discuss.

Re: [WIRELESS-LAN] EIGRP equal cost load balancing over wireless bridges

[Kenny, Eric]

Hi Mark,

Have you tried both links individually?  If they are different vendors, one 
might be handling MTU differently, which might upset your CAPWAP tunnels.  Just 
a thought.

Thanks,
--- 
Eric Kenny
Network Architect
Harvard University IT
---

> On Feb 27, 2018, at 7:45 PM, Mark Duling  wrote:
> 
> Hello all,
> 
> Is anyone running two pairs of inexpensive wireless bridges–say ubiquity 
> AirFiber–and using EIGRP equal cost load balancing over them? It seems to me 
> that should be an inexpensive way to support reasonably high bandwidth 
> building over redundant links.
> 
> I ask because a while back as a test I changed a remote building with two 
> pairs of wireless transparent bridges (one acting as primary / other acting 
> as backup, each pair a different vendor but both wired interfaces 100 Mb) 
> from routing over the primary to use equal cost load balancing over both. 
> After that we'd occasionally see our Cisco capwap APs disassociate and 
> re-associate from our WLCs on campus where they weren't doing that when data 
> was routed over a single link.
> 
> Does anyone know why that might happen with ECLB? Or in any case, is anyone 
> successfully using a dual link wireless bridge setup with both links active? 
> Thanks
> 
> Mark - Biola IT Operations
> ** Participation and subscription information for this EDUCAUSE 
> Constituent Group discussion list can be found at 
> http://www.educause.edu/discuss.
> 


**
Participation and subscription information for this EDUCAUSE Constituent Group 
discussion list can be found at http://www.educause.edu/discuss.

RE: [WIRELESS-LAN] EIGRP equal cost load balancing over wireless bridges

[Hales, David]

I can’t think of any interaction between EIGRP ECLB and the WLC that would 
cause this issue without some other additional factor.  Are you running per 
destination or per packet load balancing?  If it’s per destination, traffic 
back to a controller might be on one link that might be dropping.  Is there 
anything in the logging on the wireless bridges that might indicate that one of 
the links is flapping?  Running per packet has the possibility of causing out 
of order packet delivery.  I’m not familiar enough with the WLC traffic to know 
if that would cause any issues.  Someone else might be able to speak to that?

David Hales
Network Systems Administrator
Information Technology Services
1010 N. Peachtree
Clement Hall 117
Cookeville, TN 38505
P 931-372-3983
F 931-372-6130
E dha...@tntech.edu
www.tntech.edu/its
[Tennessee Tech Logo]
[TTU Facebook]  [TTU Twitter]  
 [TTU Instagram]  
 [TTU Youtube]  
 [TTU Pintrest] 


From: The EDUCAUSE Wireless Issues Constituent Group Listserv 
[mailto:WIRELESS-LAN@LISTSERV.EDUCAUSE.EDU] On Behalf Of Mark Duling
Sent: Tuesday, February 27, 2018 6:45 PM
To: WIRELESS-LAN@LISTSERV.EDUCAUSE.EDU
Subject: [WIRELESS-LAN] EIGRP equal cost load balancing over wireless bridges

Hello all,

Is anyone running two pairs of inexpensive wireless bridges–say ubiquity 
AirFiber–and using EIGRP equal cost load balancing over them? It seems to me 
that should be an inexpensive way to support reasonably high bandwidth building 
over redundant links.

I ask because a while back as a test I changed a remote building with two pairs 
of wireless transparent bridges (one acting as primary / other acting as 
backup, each pair a different vendor but both wired interfaces 100 Mb) from 
routing over the primary to use equal cost load balancing over both. After that 
we'd occasionally see our Cisco capwap APs disassociate and re-associate from 
our WLCs on campus where they weren't doing that when data was routed over a 
single link.

Does anyone know why that might happen with ECLB? Or in any case, is anyone 
successfully using a dual link wireless bridge setup with both links active? 
Thanks

Mark - Biola IT Operations
** Participation and subscription information for this EDUCAUSE 
Constituent Group discussion list can be found at 
http://www.educause.edu/discuss.

**
Participation and subscription information for this EDUCAUSE Constituent Group 
discussion list can be found at http://www.educause.edu/discuss.

RE: [WIRELESS-LAN] Offline/Spare Gear Inventory Size

[Ian Lyons]

Fascinating.  While we are (we are engineers after all) beating the horse 
senseless, this is interesting to see how others arrive at their decisions.

I have always worked for schools where the budget is X.  Spend it on spares or 
get gear out the students, it is on us to make that decision.  I have always 
erred on the side of less spares more for the students….

However, don’t forget the basics –environmental variables.

Before I moved to Florida, I would have maybe 2-3 switches (we had 200 switches 
in use) and ~5 AP’s for the 1200 in use fail.  I never ran out of spares and 
had an exposure…

Until I moved to Florida.

Now, a good storm can blow up 2-3 switches in a day and 6-10 ap’s.  Worst storm 
to date was 2 chassis blades, 12 switches and 32 AP’s.

So all things being equal never discount the environment you live in and your 
mileage may vary, you know your environment better than anyone (or will know it 
after a year ;)   )

Also don’t forget the product, when I was a Cisco wifi shop (in Florida) during 
the Wave 2 refresh (which people are still talking about) I had a huge failure 
rate.  Since I went to Aruba I am down to (lightening hits aside) about 1-2  
failures a year out of 1500 aps.  Conversely, I do not lose Cisco switches at 
all (lightening hits aside)

Ian.

From: The EDUCAUSE Wireless Issues Constituent Group Listserv 
[mailto:WIRELESS-LAN@LISTSERV.EDUCAUSE.EDU] On Behalf Of Frans Panken
Sent: Wednesday, February 28, 2018 3:14 AM
To: WIRELESS-LAN@LISTSERV.EDUCAUSE.EDU
Subject: Re: [WIRELESS-LAN] Offline/Spare Gear Inventory Size


I have a background in mathematics and addressed this question scientifically. 
The number of spares depend on the mean-time-between-failures of an AP 
(provided by vendors; e.g., Cisco’s 3702 APs have a MTBF of 325000 hours), the 
number of days it takes to order a new AP (k) and the risk you are accept to 
take that you have no spare AP available. If n is the spare factor (1 spare AP 
for every n active APs) and [cid:image005.png@01D3B072.CCF5AF90]  is the risk 
you accept that you have no spare AP available, this number can be computed as 
follows:

[cid:image006.png@01D3B072.CCF5AF90]

Here “ln” is the natural log, ln(2,71828…) = 1. So, in the case of Cisco AP3700 
and you accept an availability of 0.95 (5% of the time you have n spare AP) and 
it takes 6 days to order a new AP, you need one spare on every 116 APs.
-Frans

P.S. I was reluctant to spam you with the derivation; please send me an e-mail 
if you are interested in how derived this formula.



From: The EDUCAUSE Wireless Issues Constituent Group Listserv 
mailto:WIRELESS-LAN@LISTSERV.EDUCAUSE.EDU>> 
on behalf of Greg Briggs mailto:brigg...@plu.edu>>
Reply-To: The EDUCAUSE Wireless Issues Constituent Group Listserv 
mailto:WIRELESS-LAN@LISTSERV.EDUCAUSE.EDU>>
Date: Wednesday, 28 February 2018 at 00:06
To: 
"WIRELESS-LAN@LISTSERV.EDUCAUSE.EDU" 
mailto:WIRELESS-LAN@LISTSERV.EDUCAUSE.EDU>>
Subject: Re: [WIRELESS-LAN] Offline/Spare Gear Inventory Size

I solve replacement items (of many kinds) with a formula that works for small 
and large deployments.  It works for everything from switches to toilet paper 
so I have used it for more than my day job.  It can and can be expressed in a 
spreadsheet like this "=CEILING(N1*M1,1)" where N is a column with the 
calculated need and M is a rate at which you need spares, or it is consumed 
before your next purchase.  The formula is sometimes calculated on a 
spreadsheet and sometimes it is just a rough estimate in my head.  Here are 
some example starting values for the M column.  Optionally you can add a field 
for the ceiling value to reflect things that are only available in quantities 
greater than one.  For example 10 for hotdogs and 8 for hotdog buns.  The 
formalization of this formula is adapted from one my manager used, so credit is 
due to David Allen.


  *   1.1 for something like a new line of APs you don't already have a 
deployment of.  This allows for 10 spares in a deployment of 100 in case some 
arrive DOA or I find a flaw in my plan after the order/ post install.  I would 
be more conservative if I didn't think I was ever going to need that model of 
APs elsewhere, and just take the heat if I end up short.
  *   1.01 for the subsequent deployment of APs if spares are already on hand.
  *   1.1 for something that is mission critical but we only have a few of.
  *   1 for expensive things that have a high availability feature and are 
under a reasonably quick turnaround service/replacement contract.
  *   1.2 for items that we ran out of quickly last time we made an annual 
purchase.
  *   1.05 for inexpensive things that would save some time to have spares of, 
but are only a minor inconvenience if you run out.
Modify the value for subsequent orders based on current inventory or if you 
find that the failure (or consumption) rate is higher than expected.  Another 
reason to modify the value is 

Re: [WIRELESS-LAN] Air Time Fairness

[Samuel Clements]

One gotcha is the following:

Table 6 Key Features Not Supported in Cisco Aironet 1800i, 1810 OEAP,
1810W, 1815, 1830, 1850, 2800, and 3800 Series APs
Cisco Air Time Fairness (ATF)

Be warned that, if you're using the wave 2 platforms, ATF is not supported
according to the latest WLC version 8.6 release notes at:
https://www.cisco.com/c/en/us/td/docs/wireless/controller/release/notes/crn86.html

  -Sam


On Wed, Feb 28, 2018 at 1:15 AM, Jason Cook 
wrote:

> We are Cisco shop on 8.2.164 and would potentially move to 8.5 July. (8510
> in HA, peak 15k clients)
>
>
>
> Open to thoughts from other vendors as well 😊
>
>
>
> Does anyone know of any gotya’s when enabling ATF with Client Fair
> Sharing? Capacity is one that comes to mind with AVC having clearly caused
> plenty of issues under load.
>
> We are starting on our dev controller 5508’s in HA, so can play with it
> easily enough. But AVC in dev was fine also, hard to test capacity there
> since creating the load is a challenge.
>
>
>
>
>
> --
>
> Jason Cook
>
> Information Technology and Digital Services
>
> The University of Adelaide, AUSTRALIA 5005
>
> Ph: +61 8 8313 4800 <+61%208%208313%204800>
>
> e-mail: jason.c...@adelaide.edu.au
>
>
>
> CRICOS Provider Number 00123M
>
> ---
>
> This email message is intended only for the addressee(s) and contains
> information which may be confidential and/or copyright.  If you are not the
> intended recipient please do not read, save, forward, disclose, or copy the
> contents of this email. If this email has been sent to you in error, please
> notify the sender by reply email and delete this email and any copies or
> links to this email completely and immediately from your system.  No
> representation is made that this email is free of viruses.  Virus scanning
> is recommended and is the responsibility of the recipient.
>
>
> ** Participation and subscription information for this EDUCAUSE
> Constituent Group discussion list can be found at http://www.educause.edu/
> discuss.
>
>

**
Participation and subscription information for this EDUCAUSE Constituent Group 
discussion list can be found at http://www.educause.edu/discuss.

RE: Offline/Spare Gear Inventory Size

[Osborne, Bruce W (Network Operations)]

For Aruba shops, I  somewhat disagree with the last 3 points.

2% is way overkill for spares just sitting on the shelf.

Although you can cycle stock, there is no need to do so. We have had very few 
DOA Aruba APs and they do not deteriorate with age

There is absolutely no need to pull out your spare & upgrade them unless you 
expect very urgent replacements. It does not take very long for am AP to 
upgrade the code from the controller.

My advice may not apply to non-Aruba shops. YMMV.

Bruce Osborne
Senior Network Engineer
Network Operations - Wireless
 (434) 592-4229
LIBERTY UNIVERSITY
Training Champions for Christ since 1971

From: Jeffrey D. Sessler [mailto:j...@scrippscollege.edu]
Sent: Tuesday, February 27, 2018 11:23 AM
Subject: Re: Offline/Spare Gear Inventory Size


  *   Look at the turn-around time for warranty replacement. The free 
limited-lifetime may take longer than if the AP is under an extended contract.
  *   Evaluate your deployment plan. If your deployment is coverage-based, 
where the loss of a single AP could be devastating to clients, then keep more 
spares. If you have a dense deployment where the loss of one or more APs is of 
little consequence, keep less.
  *   Spares are technology collecting dust with the same life-cycle as those 
in production. If you have 5000 APs and spare 2%, that’s 100 APs that would 
likely cover a moderately sized building, and provide a lot of in-fill.
  *   If you keep spares, make sure to cycle them into production i.e. always 
install them into a new project, and put new APs back on the spare shelf.
  *   When you upgrade controller code, pull those spares out and let them 
upgrade too, then test that they still work.

Jeff

From: 
"wireless-lan@listserv.educause.edu" 
mailto:WIRELESS-LAN@LISTSERV.EDUCAUSE.EDU>> 
on behalf of "Trinklein, Jason R" 
mailto:trinkle...@cofc.edu>>
Reply-To: 
"wireless-lan@listserv.educause.edu" 
mailto:WIRELESS-LAN@LISTSERV.EDUCAUSE.EDU>>
Date: Monday, February 26, 2018 at 10:21 AM
To: 
"wireless-lan@listserv.educause.edu" 
mailto:WIRELESS-LAN@LISTSERV.EDUCAUSE.EDU>>
Subject: [WIRELESS-LAN] Offline/Spare Gear Inventory Size

Hi All,

I’m curious to know the size of your spare gear inventories. Do you keep a 
percentage of each model of AP in inventory, and what is your reasoning? 
Storms? Last minute/emergency wireless coverage needs?

What percentage of your live gear do you keep as offline inventory? (100 live 
APs with 1 inventory AP = 1% offline inventory).

With Xirrus, we had an offline inventory of more than 10% of live inventory. We 
kept that inventory to cover the high failure rate of the equipment, the 
incidence of hurricanes and lightning strikes in our area, the broad range of 
AP models on campus, and last minute large events in low coverage areas.

We are evaluating the minimum offline inventory for our new Aruba gear as we 
finish up the vendor switch. I have been thinking 1-2%, but I want to see what 
you guys do first, and why.

Thank you,
--
Jason Trinklein
Wireless Engineering Manager
College of Charleston
81 St. Philip Street | Office 311D | Charleston, SC 29403
trinkle...@cofc.edu | (843) 300–8009

DID YOU KNOW? The Princeton Review selected the College of Charleston as one of 
50 schools focused on providing students with practical experiences that take 
their academics to the next level.
** Participation and subscription information for this EDUCAUSE 
Constituent Group discussion list can be found at 
http://www.educause.edu/discuss.
** Participation and subscription information for this EDUCAUSE 
Constituent Group discussion list can be found at 
http://www.educause.edu/discuss.

**
Participation and subscription information for this EDUCAUSE Constituent Group 
discussion list can be found at http://www.educause.edu/discuss.

Re: [WIRELESS-LAN] Offline/Spare Gear Inventory Size

[Frans Panken]

I have a background in mathematics and addressed this question scientifically. 
The number of spares depend on the mean-time-between-failures of an AP 
(provided by vendors; e.g., Cisco’s 3702 APs have a MTBF of 325000 hours), the 
number of days it takes to order a new AP (k) and the risk you are accept to 
take that you have no spare AP available. If n is the spare factor (1 spare AP 
for every n active APs) and  is the risk you accept that you have no spare AP 
available, this number can be computed as follows:

[cid:image004.png@01D3B074.6EE80270]

Here “ln” is the natural log, ln(2,71828…) = 1. So, in the case of Cisco AP3700 
and you accept an availability of 0.95 (5% of the time you have n spare AP) and 
it takes 6 days to order a new AP, you need one spare on every 116 APs.
-Frans

P.S. I was reluctant to spam you with the derivation; please send me an e-mail 
if you are interested in how derived this formula.



From: The EDUCAUSE Wireless Issues Constituent Group Listserv 
 on behalf of Greg Briggs 
Reply-To: The EDUCAUSE Wireless Issues Constituent Group Listserv 

Date: Wednesday, 28 February 2018 at 00:06
To: "WIRELESS-LAN@LISTSERV.EDUCAUSE.EDU" 
Subject: Re: [WIRELESS-LAN] Offline/Spare Gear Inventory Size

I solve replacement items (of many kinds) with a formula that works for small 
and large deployments.  It works for everything from switches to toilet paper 
so I have used it for more than my day job.  It can and can be expressed in a 
spreadsheet like this "=CEILING(N1*M1,1)" where N is a column with the 
calculated need and M is a rate at which you need spares, or it is consumed 
before your next purchase.  The formula is sometimes calculated on a 
spreadsheet and sometimes it is just a rough estimate in my head.  Here are 
some example starting values for the M column.  Optionally you can add a field 
for the ceiling value to reflect things that are only available in quantities 
greater than one.  For example 10 for hotdogs and 8 for hotdog buns.  The 
formalization of this formula is adapted from one my manager used, so credit is 
due to David Allen.


  *   1.1 for something like a new line of APs you don't already have a 
deployment of.  This allows for 10 spares in a deployment of 100 in case some 
arrive DOA or I find a flaw in my plan after the order/ post install.  I would 
be more conservative if I didn't think I was ever going to need that model of 
APs elsewhere, and just take the heat if I end up short.
  *   1.01 for the subsequent deployment of APs if spares are already on hand.
  *   1.1 for something that is mission critical but we only have a few of.
  *   1 for expensive things that have a high availability feature and are 
under a reasonably quick turnaround service/replacement contract.
  *   1.2 for items that we ran out of quickly last time we made an annual 
purchase.
  *   1.05 for inexpensive things that would save some time to have spares of, 
but are only a minor inconvenience if you run out.
Modify the value for subsequent orders based on current inventory or if you 
find that the failure (or consumption) rate is higher than expected.  Another 
reason to modify the value is if replacements can be found locally, or if you 
can count a similar model of item that you have spare of as a replacement.  
MTBF is another variable to consider.

Using this formula also helps determine the support level because you can 
calculate what you need to make the cheaper support levels a better value with 
an acceptable and/or similar risk.

Greg Briggs
Network Manager
Pacific Lutheran University


On Tue, Feb 27, 2018 at 8:23 AM, Jeffrey D. Sessler 
mailto:j...@scrippscollege.edu>> wrote:

  *   Look at the turn-around time for warranty replacement. The free 
limited-lifetime may take longer than if the AP is under an extended contract.
  *   Evaluate your deployment plan. If your deployment is coverage-based, 
where the loss of a single AP could be devastating to clients, then keep more 
spares. If you have a dense deployment where the loss of one or more APs is of 
little consequence, keep less.
  *   Spares are technology collecting dust with the same life-cycle as those 
in production. If you have 5000 APs and spare 2%, that’s 100 APs that would 
likely cover a moderately sized building, and provide a lot of in-fill.
  *   If you keep spares, make sure to cycle them into production i.e. always 
install them into a new project, and put new APs back on the spare shelf.
  *   When you upgrade controller code, pull those spares out and let them 
upgrade too, then test that they still work.

Jeff

From: 
"wireless-lan@listserv.educause.edu" 
mailto:WIRELESS-LAN@LISTSERV.EDUCAUSE.EDU>> 
on behalf of "Trinklein, Jason R" 
mailto:trinkle...@cofc.edu>>
Reply-To: 
"wireless-lan@listserv.educause.edu" 
mailto:WIRELESS-LAN@LISTSERV.EDUCAUSE.EDU>>
Date: Monday, February 26, 2018 at 10:21 AM
To: 
"wir