Mailing List Archive

Hi Robert, I am forwarding this into the product team at Brocade. It does
sound disconcerting, but what I will say is that new software will continue
to be released for the ICX6xxx series, it will be as patch or maintenance
releases. I believe the reason for this is that there are some major
features being added in the next couple of releases, and the older hardware
will not support some of these. Over the past few versions, some specific
minor features have been ported or back-rev'd into certain patch releases
as well, so I wouldn't view this as the end of the line in any way for the
ICX 6xxx product. There are some unique advantages in the newer 7xxx
hardware, but the 6610s are still very versatile platforms.

Mike

On Fri, Dec 18, 2015 at 7:00 PM, Robert Hass <robhass@gmail.com> wrote:

> Hi
> Just read release notes for IronWare 8.0.40.
>
> Most of products are NOT supported starting this release. I'm very
> negative supprised.
> It's very very dirty move from Brocade to existing customers as we have
> 400+ switches.
>
> For example compact switches like ICX6430-C12 released 2.5 year ago will
> NOT receive new software, a little disaster from customer point of view ?
>
> Just few me 3 cents, as now I'm regret that we invested in Brocade.
> Probably will not buy anything from them anymore after this action. Anybody
> has same point of view like me ?
>
> Rob
>
>
> _______________________________________________
> foundry-nsp mailing list
> foundry-nsp@puck.nether.net
> http://puck.nether.net/mailman/listinfo/foundry-nsp
>

Dear Mike,

Let me add this to the mix as it seems to also concern the FCX family :

http://community.brocade.com/t5/Ethernet-Switches-Routers/Most-of-Brocade-switches-unsupported-in-last-IronWare-8-0-40/m-p/83986#U83986

Best regards.



2016-01-05 1:57 GMT+01:00 Mike Allen <mkallen@gmail.com>:

> Hi Robert, I am forwarding this into the product team at Brocade. It does
> sound disconcerting, but what I will say is that new software will continue
> to be released for the ICX6xxx series, it will be as patch or maintenance
> releases. I believe the reason for this is that there are some major
> features being added in the next couple of releases, and the older hardware
> will not support some of these. Over the past few versions, some specific
> minor features have been ported or back-rev'd into certain patch releases
> as well, so I wouldn't view this as the end of the line in any way for the
> ICX 6xxx product. There are some unique advantages in the newer 7xxx
> hardware, but the 6610s are still very versatile platforms.
>
> Mike
>
> On Fri, Dec 18, 2015 at 7:00 PM, Robert Hass <robhass@gmail.com> wrote:
>
>> Hi
>> Just read release notes for IronWare 8.0.40.
>>
>> Most of products are NOT supported starting this release. I'm very
>> negative supprised.
>> It's very very dirty move from Brocade to existing customers as we have
>> 400+ switches.
>>
>> For example compact switches like ICX6430-C12 released 2.5 year ago will
>> NOT receive new software, a little disaster from customer point of view ?
>>
>> Just few me 3 cents, as now I'm regret that we invested in Brocade.
>> Probably will not buy anything from them anymore after this action. Anybody
>> has same point of view like me ?
>>
>> Rob
>>
>>
>> _______________________________________________
>> foundry-nsp mailing list
>> foundry-nsp@puck.nether.net
>> http://puck.nether.net/mailman/listinfo/foundry-nsp
>>
>
>
> _______________________________________________
> foundry-nsp mailing list
> foundry-nsp@puck.nether.net
> http://puck.nether.net/mailman/listinfo/foundry-nsp
>



--
Youssef BENGELLOUN-ZAHR

On Tue, Jan 5, 2016 at 1:57 AM, Mike Allen <mkallen@gmail.com> wrote:

> Hi Robert, I am forwarding this into the product team at Brocade. It does
> sound disconcerting, but what I will say is that new software will continue
> to be released for the ICX6xxx series, it will be as patch or maintenance
> releases. I believe the reason for this is that there are some major
> features being added in the next couple of releases, and the older hardware
> will not support some of these. Over the past few versions, some specific
> minor features have been ported or back-rev'd into certain patch releases
> as well, so I wouldn't view this as the end of the line in any way for the
> ICX 6xxx product. There are some unique advantages in the newer 7xxx
> hardware, but the 6610s are still very versatile platforms.
>

Hi Mike

But from my point of view Patch & Maintenance means dead product. Let me do
quick explanation - from perspective of existing customer running 300+
ICX'es 6430/6450/6610/6650 as Patch & Maintenance means simply that any new
even software features (I picked up two very usefull features just for give
you example) will not appear in switches which we're using - eg. in 8.0.30
build.

Examples of new features from 8.0.40 which we probably never see in 8.0.30:
http://www.brocade.com/content/html/en/configuration-guide/fastiron-08040-managementguide/GUID-CE006189-671C-4623-B49E-F867FED75CEB.html
http://www.brocade.com/content/html/en/command-reference-guide/fastiron-08040-commandref/GUID-9C96575B-639F-411A-9D09-7E5E455F5498.html

Explanation for vendor is really simple. Integrating these features in
8.0.30 or even release 8.0.40 without new hardware features (ERSPAN, SPX)
costs because software teams receiving payrolls. But after that probably we
will not buy any new Brocade switch again. Imagine your potential profit
loss (due we will not buy anything new from Brocade) VS cost of software
teams. Maybe Brocade is not interested spending money for dead products as
they have new products - 7250/7450/7750 series. But it not just cost - it's
about customer satisfaction and vendor trust.

From my perspective as former programmer I don't see what is problem with
integrating some software features into older software release ? Is it
really so big effort and cost ? I always receiving feedback from PMs that
becouse product testing cost a lot. This never convince me - as saying that
testing cost a lot is pure bullshit as I hit a lot of times in bugs for new
features/platforms which simply means that money for testing if mostly
wasted.

Just want to add it never happen to Cisco or even HP Procurve products -
old switches are still receiving software features.
Asked my friend regarding EX series in JNPR - also old EX still receiving
new functions.

Rob

Not having the remove-vlan feature available in previous releases, despite it being the #1 complaint I get from customers (“why does changing VLANs suck so bad?”, with why does the lacp configuration suck so bad being a close #2), is enough to make you want to spit.

-Brant.

> On Jan 5, 2016, at 3:10 AM, Robert Hass <robhass@gmail.com> wrote:
>
> On Tue, Jan 5, 2016 at 1:57 AM, Mike Allen <mkallen@gmail.com <mailto:mkallen@gmail.com>> wrote:
> Hi Robert, I am forwarding this into the product team at Brocade. It does sound disconcerting, but what I will say is that new software will continue to be released for the ICX6xxx series, it will be as patch or maintenance releases. I believe the reason for this is that there are some major features being added in the next couple of releases, and the older hardware will not support some of these. Over the past few versions, some specific minor features have been ported or back-rev'd into certain patch releases as well, so I wouldn't view this as the end of the line in any way for the ICX 6xxx product. There are some unique advantages in the newer 7xxx hardware, but the 6610s are still very versatile platforms.
>
> Hi Mike
>
> But from my point of view Patch & Maintenance means dead product. Let me do quick explanation - from perspective of existing customer running 300+ ICX'es 6430/6450/6610/6650 as Patch & Maintenance means simply that any new even software features (I picked up two very usefull features just for give you example) will not appear in switches which we're using - eg. in 8.0.30 build.
>
> Examples of new features from 8.0.40 which we probably never see in 8.0.30:
> http://www.brocade.com/content/html/en/configuration-guide/fastiron-08040-managementguide/GUID-CE006189-671C-4623-B49E-F867FED75CEB.html <http://www.brocade.com/content/html/en/configuration-guide/fastiron-08040-managementguide/GUID-CE006189-671C-4623-B49E-F867FED75CEB.html>
> http://www.brocade.com/content/html/en/command-reference-guide/fastiron-08040-commandref/GUID-9C96575B-639F-411A-9D09-7E5E455F5498.html <http://www.brocade.com/content/html/en/command-reference-guide/fastiron-08040-commandref/GUID-9C96575B-639F-411A-9D09-7E5E455F5498.html>
>
> Explanation for vendor is really simple. Integrating these features in 8.0.30 or even release 8.0.40 without new hardware features (ERSPAN, SPX) costs because software teams receiving payrolls. But after that probably we will not buy any new Brocade switch again. Imagine your potential profit loss (due we will not buy anything new from Brocade) VS cost of software teams. Maybe Brocade is not interested spending money for dead products as they have new products - 7250/7450/7750 series. But it not just cost - it's about customer satisfaction and vendor trust.
>
> From my perspective as former programmer I don't see what is problem with integrating some software features into older software release ? Is it really so big effort and cost ? I always receiving feedback from PMs that becouse product testing cost a lot. This never convince me - as saying that testing cost a lot is pure bullshit as I hit a lot of times in bugs for new features/platforms which simply means that money for testing if mostly wasted.
>
> Just want to add it never happen to Cisco or even HP Procurve products - old switches are still receiving software features.
> Asked my friend regarding EX series in JNPR - also old EX still receiving new functions.
>
> Rob
> _______________________________________________
> foundry-nsp mailing list
> foundry-nsp@puck.nether.net
> http://puck.nether.net/mailman/listinfo/foundry-nsp

Regarding 'remove-vlan' - To be honest we was pinging our Brocade partner
for that kind of feature. Imagine removing 600 VLANs from one port - we
doing 'perl scripts + copy tftp running-config' as workaround. And as I
know Partner was pinging proper SE / PM for integrate feature enhancement.
Hearing there was a lot of other engineers begging for that kind of feature.

And and the end we can forget about that important command as we're ICX
6xxx series.

Exacly same story about 'show CLI history'. But here partner sends us
workaround command 'dm history' working on current platforms.

Robert, et al, I would urge you to lean on your local SE resource, we all
advocate for our customers back into corporate. In the end, engineering
resources are finite (we are hiring software engineers as fast as we can
find them) and the company allocates time accordingly. All that said, we
do add features into patch releases similar to the vendors, to address
customer requests and needs. For example, here is a snippet of the release
notes for enhancements added in 8.0.30b that did not exist in prior
releases:

*Enhancements in FastIron 08.0.30b *

Brocade FastIron Release 08.0.30b introduces serveral new features and
enahcncement.

ï‚· Flexible Authentication enhancement

o Additional RADIUS attribute support for Dynamic VLAN assignments

o Dynamic Tagged VLAN assignements not limited to Voice VLANs

o Support for single and multiple untagged VLANs per port is configurable

ï‚· Stacking enhancements

o ICX 7750-48C and ICX 7750-48F devices support stacking distances of 10 Km
using LR4 fiber optic cables attached to ports 1/2/5 and 1/2/6. Manual
trunk configuration using port 1/2/1 or 1/2/4 as a lead default stacking
port is required.

ï‚· LAG symmetric load balancing

o Sometimes DPI devices and firewalls are installed as a bump in the wire
deployment on certain child links of a LAG. In such a case symmetrical
hashing is very important for LAG interfaces. This allows the reverse flow
of traffic to be directed through the same child link on the LAG and is
bound to flow through the same DPI device. This enables proper accounting
on the DPI of the traffic in both the forward and reverse flows. The same
is true for firewall devices as well so they could filter out unwanted
traffic in both the directions.

ï‚· LAG Scaling

o In FastIron 08.0.30b, the number of LAGs supported on each ICX 7250, ICX
7450, or ICX 7750 increases to 256. When you downgrade from FastIron
08.0.30b, only the first 128 LAGs are deployed.The remaining LAGs are not
deployed, and related configuration is lost.

ï‚· DHCP snooping, DAI, and IP source guard over LAG

o DHCPv4 snooping, Dynamic ARP inspection and IP source guard are supported
over LAG. DHCPv4 snooping, Dynamic ARP inspection and IP source guard were
previously supported features and in 8.0.30b were supported over LAG as
well.

ï‚· Delay time in notifying VE down event

o When all the ports in the VLAN go into an inactive state (for example,
the non-forwarding state), the device notifies the Layer 3 protocols of the
VE down event only after the configured timer expires. Once the timer
expires, the device checks if any of the ports is in the forwarding state.
If no ports are in the forwarding state, the device notifies the Layer 3
protocols of the VE down event. If any of the ports is in the forwarding
state, the device ignores the down event.

So the shift is really how we add features. Significant new features (such
as the SPX or ERSPAN you mentioned) may depend on new hardware, and would
be implemented in what we consider a major release. For the features you
mention with the simplified vlan management and command history, those look
to be great candidates to port from 8.0.40 back to previous versions of
code. The best way to make that happen is to be vocal to your partner and
Brocade SE, and let them champion those features internally.

Thanks,
Mike

On Tue, Jan 5, 2016 at 3:06 PM, Robert Hass <robhass@gmail.com> wrote:

> Regarding 'remove-vlan' - To be honest we was pinging our Brocade partner
> for that kind of feature. Imagine removing 600 VLANs from one port - we
> doing 'perl scripts + copy tftp running-config' as workaround. And as I
> know Partner was pinging proper SE / PM for integrate feature enhancement.
> Hearing there was a lot of other engineers begging for that kind of
> feature.
>
> And and the end we can forget about that important command as we're ICX
> 6xxx series.
>
> Exacly same story about 'show CLI history'. But here partner sends us
> workaround command 'dm history' working on current platforms.
>
>

Hi

Well, from customer perspective It's hard to hear that engineering
resources are finite. Especially being ICX6650 or ICX6450-C12-PD customer
(we own a lot of both) - both products are just 2 years and 3 months old.
I'd say network device after 2 years since FCS are finally really stable
for critical environments - most of bugs are finally corrected ;-).

Just my 2 cents. I hope you will ports new pure-SW features to 8.0.30 next
builds.

Also ICX7750 price perspective cannot replace ICX6650 when you need maximum
20-26 10GE Ports.

Below are created based on ICX release notes.

Release date Product
----------------------------------------------------------------------
09-30-2013 ICX6450-C12-PD
09-20-2013 ICX6650
06-25-2013 ICX6430-C12, SX800-Gen3
03-12-2012 ICX6430, ICX6450
10-13-2011 ICX6610

Rob

On Thu, Jan 7, 2016 at 7:21 PM, Mike Allen <mkallen@gmail.com> wrote:

> Robert, et al, I would urge you to lean on your local SE resource, we all
> advocate for our customers back into corporate. In the end, engineering
> resources are finite (we are hiring software engineers as fast as we can
> find them) and the company allocates time accordingly. All that said, we
> do add features into patch releases similar to the vendors, to address
> customer requests and needs. For example, here is a snippet of the release
> notes for enhancements added in 8.0.30b that did not exist in prior
> releases:
>
> *Enhancements in FastIron 08.0.30b *
>
> Brocade FastIron Release 08.0.30b introduces serveral new features and
> enahcncement.
>
> ï‚· Flexible Authentication enhancement
>
> o Additional RADIUS attribute support for Dynamic VLAN assignments
>
> o Dynamic Tagged VLAN assignements not limited to Voice VLANs
>
> o Support for single and multiple untagged VLANs per port is configurable
>
> ï‚· Stacking enhancements
>
> o ICX 7750-48C and ICX 7750-48F devices support stacking distances of 10
> Km using LR4 fiber optic cables attached to ports 1/2/5 and 1/2/6. Manual
> trunk configuration using port 1/2/1 or 1/2/4 as a lead default stacking
> port is required.
>
> ï‚· LAG symmetric load balancing
>
> o Sometimes DPI devices and firewalls are installed as a bump in the wire
> deployment on certain child links of a LAG. In such a case symmetrical
> hashing is very important for LAG interfaces. This allows the reverse flow
> of traffic to be directed through the same child link on the LAG and is
> bound to flow through the same DPI device. This enables proper accounting
> on the DPI of the traffic in both the forward and reverse flows. The same
> is true for firewall devices as well so they could filter out unwanted
> traffic in both the directions.
>
> ï‚· LAG Scaling
>
> o In FastIron 08.0.30b, the number of LAGs supported on each ICX 7250,
> ICX 7450, or ICX 7750 increases to 256. When you downgrade from FastIron
> 08.0.30b, only the first 128 LAGs are deployed.The remaining LAGs are not
> deployed, and related configuration is lost.
>
> ï‚· DHCP snooping, DAI, and IP source guard over LAG
>
> o DHCPv4 snooping, Dynamic ARP inspection and IP source guard are
> supported over LAG. DHCPv4 snooping, Dynamic ARP inspection and IP source
> guard were previously supported features and in 8.0.30b were supported over
> LAG as well.
>
> ï‚· Delay time in notifying VE down event
>
> o When all the ports in the VLAN go into an inactive state (for example,
> the non-forwarding state), the device notifies the Layer 3 protocols of the
> VE down event only after the configured timer expires. Once the timer
> expires, the device checks if any of the ports is in the forwarding state.
> If no ports are in the forwarding state, the device notifies the Layer 3
> protocols of the VE down event. If any of the ports is in the forwarding
> state, the device ignores the down event.
>
> So the shift is really how we add features. Significant new features
> (such as the SPX or ERSPAN you mentioned) may depend on new hardware, and
> would be implemented in what we consider a major release. For the features
> you mention with the simplified vlan management and command history, those
> look to be great candidates to port from 8.0.40 back to previous versions
> of code. The best way to make that happen is to be vocal to your partner
> and Brocade SE, and let them champion those features internally.
>
> Thanks,
> Mike
>
> On Tue, Jan 5, 2016 at 3:06 PM, Robert Hass <robhass@gmail.com> wrote:
>
>> Regarding 'remove-vlan' - To be honest we was pinging our Brocade partner
>> for that kind of feature. Imagine removing 600 VLANs from one port - we
>> doing 'perl scripts + copy tftp running-config' as workaround. And as I
>> know Partner was pinging proper SE / PM for integrate feature enhancement.
>> Hearing there was a lot of other engineers begging for that kind of
>> feature.
>>
>> And and the end we can forget about that important command as we're ICX
>> 6xxx series.
>>
>> Exacly same story about 'show CLI history'. But here partner sends us
>> workaround command 'dm history' working on current platforms.
>>
>>
>