So… I’m currently working on a project to deploy a new UCM cluster. One thing is really just not clicking for me though as I try to read and understand it better. DSP’s and how they are configured for conference resources with UCM.

Yes I know about the Cisco DSP calculator. Yes I have two books that cover the configuration around the topic both on the gateways and within UCM. I have “Configuring Cisco Unified Communications Manager and Unity Connection, Second Edition” and “Cisco Voice Gateways and Gatekeepers”.

I’ve also been reading what the SRND’s have to say about this, which actually cleared up one fact but completely messed me up on calculations. The SRND’s are either wrong, inconsistent or something.. but things don’t match up.

Here is what I configured and what I see on a 3845 that I am using to get a better idea of things with.

BTW – I am running the following IOS version: Version 12.4(25d) – c3845-ipvoicek9-mz.124-25d.bin

So I have a 3845 with two PVDM2-48′s. Each PVDM2-48 has 3 DSP’s on it. 1 DSP has 16 “channels”, or 240 credits. A conf. call uses 15mips (G.711) / 240 credits = 16.

NAME: “PVDMII DSP SIMM with three DSPs”, DESCR: “PVDMII DSP SIMM with three DSPs”
PID: PVDM2-48 , VID: V01 , SN: FOC10482P0T

NAME: “PVDMII DSP SIMM with three DSPs”, DESCR: “PVDMII DSP SIMM with three DSPs”
PID: PVDM2-48 , VID: V01 , SN: FOC10153EB8

Using G.711 requires fewer resources and thus you can get more calls out of it.

According to the SRND, 1 “session” is 1 conference call. 1 conference call can have 8 participants. 1 session = 1 conference call. Even if you only have 3 people on the conference call the system allocates all 8 channels to that conference session.

I decide to use 1 DSP and see what it would let me configure as the max sessions:

dspfarm profile 1 conference
codec g711ulaw
codec g711alaw
maximum sessions 2
associate application SCCP

So far this makes sense I think. 1 DSP = 16 channels, 1 conference call uses 8 channels so your max resources would be 2 sessions / conferences.

VGWY_3845_1#sh voice dsp group all
–output removed–
dsp 7:
State: UP, firmware: 1.1.1
Max signal/voice channel: 16/16
Max credits: 240
Group: FLEX_GROUP_CONF, complexity: CONFERENCE
Shared credits: 0, reserved credits: 60
Conference session: 8
Credits used: 0

VGWY_3845_1#sh dspfarm profile 1
Dspfarm Profile Configuration

Profile ID = 1, Service = CONFERENCING, Resource ID = 1
Profile Description :
Profile Admin State : UP
Profile Operation State : ACTIVE IN PROGRESS
Application : SCCP Status : ASSOCIATION IN PROGRESS
Resource Provider : FLEX_DSPRM Status : UP
Number of Resource Configured : 2
Number of Resource Available : 2
Codec Configuration
Codec : g711ulaw, Maximum Packetization Period : 30 , Transcoder: Not Required
Codec : g711alaw, Maximum Packetization Period : 30 , Transcoder: Not Required

So far so good I think. These numbers make sense to me.

HOWEVER… this is not what many say is the max you should get.

For instance, here is one thing I have read:

“Flex mode allows for oversubscription of DSPs. CNF takes a full DSP (PVDM2), either 2 or 8 sessions (mixed or G711 only)”

As you can see above the CONF config is in Flex mode, its G.711 only, I allocated a full DSP from a PVDM2. I can only configure 2 max sessions?

Here is another interesting point from two different SRND’s.. neither of which make sense to me and add up.

7.x SRND – http://www.cisco.com/en/US/docs/voice_ip_comm/cucm/srnd/7x/media.html#wp1046210

“If G.711 conferences are configured, then no more than 6 DSPs (total
of 48 conferences with 8 participants each) should be allocated per NM
because (48 * 8 ) participants = 384 streams”

How the heck do you get 48 conferences – max sessions from 6 DSPs?

4.x SRND – http://www.cisco.com/en/US/docs/voice_ip_comm/cucm/srnd/4x/42media.html#wp1079416

“If G.711 conferences are configured, then no more than 48 DSPs should
be allocated per NM because (48 * 8 ) participants = 384 streams”

Ok.. so at least this time they didn’t say 6 DSPs, but how the heck do you get 48 DSPs on one NM that has only four slots? PVDM2-64 = 4 DSPs. Four PVDM2-64′s on one NM = 16 DSPs not 48 DSPs?

What the heck am I missing here?

IPv6 huh? Yes, IPv6. It’s time people.. seriously it’s past time. If your a network or system admin / engineer you need to get started learning all you can about IPv6. No more putting it off!

There is a great new resource out there that I encourage everyone to take advantage of if your lagging behind on your IPv6 knowledge. The site was done by someone I’ve known for about 3 years now, Brandon Carroll. He’s a well known instructor and author. He has a CCIE in Security and really knows his stuff.

His site is: www.myipv6tutor.com

He currently is offering an early bird promotion where you can register for $197 until May 15th.

You get immediate access to the first three instructional videos covering:

• IPv6 Features and Benefits
• IPv6 Addressing
• IPv6 Headers

Once the program fully launches you’ll gain access to the remaining lessions:

• ICMPv6 – Why Is This Important?
• Stateless and Stateful Autoconfiguration
• Basic Configuration of IPv6 on Routers and Desktop Computers

And your registration also allows you access to all the MP3′s of each lession as well as the PDF transcripts for each lesson. You will even get some lab exercise documents to practice what you learn and a free 2-hour live web session on IPv6 Security.

There is no doubt Brandon has put a lot of work and effort into this site and the videos. After reviewing most of the items I can attest to his level of expertise on the subject as well as experience as an instructor. Honestly he’s just a great guy to know, a great instructor and he is offering a great service and value in my humble opinion.

If you need or want IPv6 training for yourself or your team I highly recommend this as a very affordable option to most other training out there today. Your able to watch the videos from the comfort of your home or office, listen to the MP3′s while you commute and work on lab scenerios when time permits.

Also, if your really serious about getting up to speed on IPv6 I would also recommend you check out Hurricane Electrics free IPv6 Tunnel Broker. You can sign up for free, and get access to setup IPv6 within your home or lab by tunneling it over IPv4 to he.net.

This will really come in handy and help to supplement your IPv6 training with www.myIPv6tutor.com

RST Reports: An RST report is a report from a receiving station on the quality and strength of the transmitted signal. Using shorthand in the form of numbers to represent the tone of a CW signal or voice transmission of a transmitting station’s signal at the receiving station’s location (QTH).

Here is what it means:

R Readability – Understanding what is said and how well. On a scale of 1 to 5, the readability of your signal with a “5″ being perfect with no difficulty. In other words the ability of the other operator to understand what you are saying. A “1″ is unreadable….a “5″ is perfectly readable.

S Strength – On a scale of 1 to 9, indicates how strong your stations signal is. A “1″ is a very faint signal. A “9″ is an extremely strong signal.

T Tone – Used for morse code signal reports. Indicates on a scale of 1 to 9 the quality of the tone of the morse code “dits and dahs”. From a “60 cycle harsh tone” a (1)…. To a “very pure tone”, a (9).

R = READABILITY
1 — Unreadable
2 — Barely readable, occasional words distinguishable
3 — Readable with considerable difficulty
4 — Readable with practically no difficulty
5 — Perfectly readable

S = SIGNAL STRENGTH
1 — Faint signals, barely perceptible
2 — Very weak signals
3 — Weak signals
4 — Fair signals
5 — Fairly good signals
6 — Good signals
7 — Moderately strong signals
8 — Strong signals
9 — Extremely strong signals

T = TONE
1 — Sixty cycle a.c. or less, very rough and broad
2 — Very rough a.c. , very harsh and broad
3 — Rough a.c. tone, rectified but not filtered
4 — Rough note, some trace of filtering
5 — Filtered rectified a.c. but strongly ripple-modulated
6 — Filtered tone, definite trace of ripple modulation
7 — Near pure tone, trace of ripple modulation
8 — Near perfect tone, slight trace of modulation
9 — Perfect tone, no trace of ripple or modulation of any kind

Cisco has been adding lots of new features and capabilities to NX-OS over the last couple of years. So much in fact that it appears the original 4GB of memory that came standard on the Nexus 7010 SUP (N7K-SUP1) now needs to be upgraded to 8GB.

The new N7K-SUP1′s from Cisco come with 8GB standard now, but those of us who have had ours for a couple of years are finding ourselves in need of an upgrade.

The 8GB upgrade requirement is based on the NX-OS version, particularly NX-OS 5.1(1) or later. It’s important to note however that not everyone planning to run NX-OS 5.1(1) or later HAS to upgrade the ram to 8GB just yet. The document Cisco Nexus 7000 Series NX-OS Release Notes, Release 5.1 explains the current situations that would require one to upgrade to 8GB. They are:

• If you are running Cisco NX-OS Release 5.1(x) and you have created more than three VDCs, or if XL mode is enabled in more than one VDC, then you must upgrade the memory to 8 GB.

• If you are running Cisco NX-OS Release 5.1(x), 8 GB of memory is required if FabricPath or FEX features are enabled in more than two VDCs of a Cisco Nexus 7000 Series system.

It is my personal opinion however that even if you do not currently match any of these requirements on your Nexus 7010 and especially if you have (AND YOU SHOULD!) dual SUPs then you should go ahead and get the memory upgrade done now before you run into a situation down the road where your running NX-OS 5.1(1) or higher and you go and create a third VDC or you try to enable FabricPath or FEX and potentially run into a memory issue.

Ok, so you decide you want or need the 8GB memory upgrade. I suggest the first thing you do is confirm that your SUPs are one’s that currently have just 4GB. You can do that with the “show system resources” command.


NXD02# show system resources
Load average: 1 minute: 0.94 5 minutes: 0.48 15 minutes: 0.35
Processes : 331 total, 2 running
CPU states : 0.5% user, 3.0% kernel, 96.5% idle
Memory usage: 4115740K total, 1606288K used, 2509452K free

As you can see the “Memory usage: 4115740K total” indicates this SUP currently has just the 4GB of memory.

Now that you have confirmed you do in fact have just 4GB of memory your going to need to order the 8 GB memory upgrade kit from Cisco, “N7K-SUP1-8GBUPG=”. Currently the list price for this is $4,000 per kit. I sure hope you have a good discount with Cisco. Oh and remember you need one kit for each Nexus 7010 SUP (N7K-SUP1)!

After placing your order with Cisco and likely waiting your 20+ day lead time you will finally be greeted with your fresh new 8GB memory upgrade kits! Yay!

So now your ready to do your upgrade. Before you do I have a couple of recommendations to make. Also its important to note that all that I am writing about is based on the fact that my particular Nexus 7010′s both have dual sups. Without the dual sups well your probably going to want and need to do a little more planning of course.

So about my recommendations. First, if you have others that have access to the Nexus 7010′s I’d suggest informing them of your intentions to do this upgrade and to not be making any changes on the systems at this time. Second, take a few minutes to go and copy your running config before you get started. Finally, just as an FYI, if your running dual sups and do everything I’ve suggested thus far you should know its pretty easy and seemingly safe to do the upgrade any time really. I know I know… scheduled maintenance, but seriously I’ve yet to have a problem doing any upgrades both software and hardware even during business hours. Yes, I LOVE the Nexus platform.

With that said I’d still recommend you backup your config, show run and copy paste into txt file. I never do any work without at least doing that.

Copy running config:
NXD02# copy run start
[########################################] 100%
Copy complete, now saving to disk (please wait)...
NXD02#

Ok, so you’ve copied the running configs, you have your txt file backups and you’re ready to get going.

Next step is check the status of the sups and redundancy:

NXD02# sh redundancy status
Redundancy mode
---------------
administrative: HA
operational: HA

This supervisor (sup-5)
-----------------------
Redundancy state: Active
Supervisor state: Active
Internal state: Active with HA standby

Other supervisor (sup-6)
------------------------
Redundancy state: Standby

Supervisor state: HA standby
Internal state: HA standby

System start time: Fri May 15 18:21:42 2009

System uptime: 695 days, 2 hours, 26 minutes, 8 seconds
Kernel uptime: 11 days, 13 hours, 1 minutes, 49 seconds
Active supervisor uptime: 11 days, 12 hours, 38 minutes, 41 seconds
NXD02#

This confirms that Sup-5 is active and that Sup-6 is your standby and is going to be the first one we upgrade.

You can also tell by looking at the indicator lights, hard to tell in the pic but the “Active” light is amber:

From the above pic you can also see the “Ejector” buttons on the SUP, at this point go ahead and unscrew the two screws and press the ejector buttons to release the levers so you can remove the SUP.

We now have a poor lonely active sup patiently waiting for it’s friend to return:

Here is the SUP with the current 4GB of memory. You can see the existing 4GB of memory in the bottom right of the photo. There are two slots for memory and we will be adding a second 4GB into the second slot right above the existing slot.

Now go ahead and get your new memory modules ready:

Now you can go ahead and insert the new memory module into the second slot. Make sure the latches click into place to secure the new memory module in place:

Once the memory module has been inserted it’s time to now do what I found to be the hardest part of the whole process, adding the darn little “8GB” sticker to the front of the SUP. Took me more time to get that little sticker on straight then anything else.

Once the hardware upgrade is done its time to reinsert the SUP back into the Nexus 7010.

Now your going to want to watch the progress of the SUP coming back online. For me I basically follow along with a few “show redundancy status” commands issued at various times as well as the console output of course.

As soon as you reinsert the SUP you will notice the first thing is it will be “At BIOS”:


NXD02# sh redundancy status
Redundancy mode
---------------
administrative: HA
operational: None

This supervisor (sup-5)
-----------------------
Redundancy state: Active
Supervisor state: Active
Internal state: Active with no standby

Other supervisor (sup-6)
------------------------
Redundancy state: At BIOS

Supervisor state: N/A
Internal state: N/A


Then the SUP is going to initialize:


NXD02# sh redundancy status
Redundancy mode
---------------
administrative: HA
operational: None

This supervisor (sup-5)
-----------------------
Redundancy state: Active
Supervisor state: Active
Internal state: Active with no standby

Other supervisor (sup-6)
------------------------
Redundancy state: Initializing

Supervisor state: N/A
Internal state: N/A


Right after that you should see “assuming your consoled in of course”:

2011 Apr 4 09:16:31 NXD02 %SYSMGR-2-ACTIVE_LOWER_MEM_THAN_STANDBY: Active supervisor in slot 5 is running with less memory than standby supervisor in slot 6.

2011 Apr 4 09:17:12 NXD02 %IDEHSD-STANDBY-2-MOUNT: slot0: online

2011 Apr 4 09:17:20 NXD02 %BOOTVAR-5-NEIGHBOR_UPDATE_AUTOCOPY: auto-copy supported by neighbor supervisor, starting...


Now the SUP is up and it will begin sync’ing with Active SUP:

NXD02# sh redundancy status
Redundancy mode
---------------
administrative: HA
operational: None

This supervisor (sup-5)
-----------------------
Redundancy state: Active
Supervisor state: Active
Internal state: Active with HA standby

Other supervisor (sup-6)
------------------------
Redundancy state: Standby

Supervisor state: HA standby
Internal state: HA synchronization in progress


Once the sync is complete the SUP should be ready in standby mode. You should see this within the console session:

2011 Apr 4 09:18:24 NXD02 %MODULE-5-STANDBY_SUP_OK: Supervisor 6 is standby

You can also confirm the status yourself:

NXD02# sh redundancy status

Redundancy mode
---------------
administrative: HA
operational: HA

This supervisor (sup-5)
-----------------------
Redundancy state: Active
Supervisor state: Active
Internal state: Active with HA standby

Other supervisor (sup-6)
------------------------
Redundancy state: Standby

Supervisor state: HA standby
Internal state: HA standby


Once things have settled and the SUP is back online and is in sync with the Active SUP your going to want to do a “system switchover” which will now make your Standby SUP (the one we just upgraded) the Active SUP.

After the switchover occurs (remember to move your console cable to the now Active SUP) you can check the system resources and confirm that you now have your 8GB of memory.


NXD02# sh system resources
Load average: 1 minute: 0.94 5 minutes: 1.49 15 minutes: 0.85
Processes : 329 total, 1 running
CPU states : 2.0% user, 3.0% kernel, 95.0% idle
Memory usage: 8254672K total, 1560576K used, 6694096K free


Well there you have it. Your Nexus 7010 is now running and active on the upgraded SUP and you’ve got your 8GB of memory installed.

Now you can move on to upgrading the second (now standby SUP) as well.

Good luck!