Intro
My last post asked for GNS3 to come clean. Some may have mistaken it for a bashing
of GNS3 which was definitely not the intention of the post. The post was intended
to get a discussion going and to see what GNS3 is planning for the future.
I’m happy to say that Stephen Guppy of GNS3 reached out to me to comment on the post.
They are listening to the community which is fantastic.
Switching
First off regarding the switching, yes it will be implemented through L2IOU.
If you expect GNS3 to develop new switches to be implemented in the new version,
this is not the case. For anyone with some knowledge of switches you must understand
how difficult this is. Switches do all the forwarding in ASICs which are very fast
and dumb chips designed to be lightning fast when forwarding traffic.
To emulate an ASIC would be very difficult if not impossible and the design of the
ASIC is not anything that Cisco would release to the public.
Of course you could implement other switches such as Arista or Huawei or anything
that has been released as a VM. GNS3 is working very hard to implement native support
for these images.
Is it legal?
GNS3 is providing a frontend to L2IOU. They can’t provide binaries or licenses. IOU
is a Cisco only application. It is up to the end users to choose if they want to use
this functionality or not. Looking from Cisco’s perspective it’s a double edged sword.
They don’t want to have something in the wild which could be looked upon as a Cisco
official image and not having the quality as expected from Cisco. On the other hand
Cisco will surely benefit from users having hands on experience on their operating
systems. In the end it’s all up to you if you choose to use such binaries.
The future of GNS3
In the beginning GNS3 was mainly used by students wanting to get certified on Cisco
equipment. Stephen pointed out that although Cisco still is an important player, GNS3
has grown a lot and many students now study Juniper and other vendors which make up
a big chunk of the users.
The aim of GNS3 in the future is to be the “CML/VIRL” of platforms. CML will be for
Cisco only although you could bridge adapters out to other networks. GNS3 wants
to integrate with all vendors to make it a multi vendor environment where you can
run all networking operating systems that are avaible as VMs. This could be XRv,
CSR1000v, N1kv, Vyatta, Arista, Juniper and so on.
End words
I hope this post has brought some clarity on how the switching in the new GNS3 will
be implemented. I wish Stephen and the rest of the GNS3 team good luck with their
new release.
“GNS3 is providing a frontend to L2IOU. They can’t provide binaries or licenses. IOU
is a Cisco only application. It is up to the end users to choose if they want to use
this functionality or not”
Hey, can you explain what he meant by “it is up to the end users to choose if they want to use this functionality or not”?
What I understand is that we could also use the Cisco L2IOU binaries with the new GNS3, if we have a license for it
excuse the ignorance….I don’t have much knowledge on L2IOUs
Unless you are Cisco employee you don’t have legal access to IOU. If you were to attain such an image and license you could use it but it’s up to the end user if you want to use these images that you don’t have the right to.
I’ve been thinking about this a bit, and I wonder if GNS3 will be able to use the release of CML to help them with IOU. Previously you were running IOS images, because Cisco released those, and it was up to the user to provide the IOS image.
But with CML, Cisco is effectively “releasing” IOL – so if I’ve paid for CML (and therefore IOL), can I then use the IOL image in GNS3? There will probably be something in the license text about not allowing it, but it’s not nearly as bad as obtaining IOL today. Not that I would ever do that, no sir…
I was bearish on the future of GNS3, but I’ve thought more about it, and I think there’s a real place for them to act as a “multi-vendor CML” – there is MASSIVE value in this. They could even hook into CML for the Cisco parts, and hook into other vendor tools as required. (Junosphere, etc).
Hi Lindsay,
Yes, I was a bit doubtful about GNS3 future but I think they might have a place to do multi vendor and not have to use several separate tools to do the same job.
My understanding of CML is that you buy the framework, no images are included. You can then buy or download images to run in CML. CSR1000v and XRv with demo license are free and I suppose that they will remain so. I don’t know how they will handle NXOS (Titanium) and IOS (vIOS). Maybe they will create demo version for these as well or they will be under contract. Regarding switching I don’t think it will be implemented from the start in CML but I think they are working on it.
Integrating L2IOU and normal IOS images will be a bit messy though regarding hardware resources but still its good road to follow. I think GNS3 team has already started to go for multi-vendor option. GNS3 has already achieved some milestones and forcing Cisco to provide tools like VIRL/CML etc.
Personally i like something that require less hardware resources. IOU is a good tool but unfortunately it’s not available to us 🙂
Good job Daniel!!
Thanks Murad,
I agree that these new images use a lot of memory and resources. I think the future of networking labs is to run it in virtualized environment. Initially you will have to buy hardware such as server, memory, storage etc which could be costly but on the other hand much cheaper than buying the real equipment.
I feel like the crowdfunding was a bait-and-switch. I donated very specifically because they stated “we can emulate switching, it’s just hard to do.” The truth is, they CAN’T. They can link to an illegal-to-use application.
If they had just stated “hey we’d like to make GNS3 better, we need money” I would have been happy to donate, because I have found GNS3 to be incredibly useful in spite of the lack of switching.
But, because we all find out, > $250k later, that the switching will be available via L2IOU, I’m just pissed and I wouldn’t donate to them again no matter what… because this was a dirty way to get funding by promising switching without explaining that you have to download an illegal product from questionable sites to use it. People on the Cisco network have been like “what’s the big deal, it’s a useful product,” yes I get that, and yes it is. But this was misleading at BEST and they freaking knew it. Shameful, guys.
Another slimy and amateurish thing that seems to have been done lately is that here is no documentation available unless you register. So unless you give up an email address you can’t even read up on whether the product can do what you want.