Dynamips, along with the control tools dynagen and GNS3, have become the one of the biggest assets to anyone studying, learning, or trying to build and test systems built with Cisco routers. For those trying to improve their Cisco skills, it means the difference between just reading the book, or attending the class, and being able to spend time practicing the feature, understanding it, and becoming fully comfortable with it. There is a huge difference between engineers who’ve read about things, and those who’ve experience using them. We’ve all met plenty of both. The importance of this hands on lab time cannot be overestimated, and while nothing beats ‘real job’ experience, lab work is second best.
The existence of dynamips has been a positive thing for Cisco, in that it (along with their clever marketing of Cisco Academy it has to be said) has helped lead to the existence of thousands of knowledgeable engineers around the world. Engineers who natively speak IOS, and know the Cisco approach to networking. This ecosystem of cisco speakers is a major factor in many firms decisions to go ‘pure cisco’. It’s simple – you can easily hire the staff to drive Cisco kit. There is no cost to Cisco due to the existence of Dynamips – there isn’t a cisco product that people don’t buy instead.
The end is Nigh
However this week marked the beginning of the end for this old friend, with the announcement of the supported hardware platforms for IOS 15.1, we see that only ISR and ISR2 routers are supported. For architectural reasons, it’s never been possible to emulate these platforms in dynamips, so we’ve reached a place where new IOS software won’t be usable in dynamips. We can run 12.4 on the existing range of platforms, and 15.0 on the 7200 platform, and thats it.
Today, it doesn’t really affect anything, unless you want to try the new hold music feature in 15.1, but over the months, as we head into 15.2 and 15.3, then the software we can run in dynamips will become too old to matter any more. Exam blueprints will move on, features we want to use will only exist in the new versions, we’ll update our production and test networks, and slowly, dynamips will become a thing of the past.
For those of us who have invested heavily in our own training and lab equipment, it won’t make a huge difference. Adding a few second hand 1841’s won’t cost a fortune, although will use a lot more power and cooling than the single dynamips box does. But for the ordinary CCNA and CCNP student, it’ll be too much to buy, too much room taken up, and too much power and noise. They can use remote labs, but they really don’t give the flexibility of your own kit.
There are great opportunities here. Cisco could release a version of IOU to the public – in the same way that F5 released the LTM VE edition recently to encourage adoption of their main product. It would be performance limited of course, to stop production use, but full featured. I don’t even mind paying a fair price for it. But guy’s – if you want to keep the geeks loyal, you need to let them know your software well.
And if there was ever an opportunity for Juniper to follow up on their ‘JunOS as a second language’ program – then this is it. They could hold out the olive branch (OK, it’s a terrible pun) to the community and either release or encourage a similar product. ‘Hey guys, come over here, you can learn our stuff much easier than Cisco’s’. There are a lot of vendors out there trying to tell us their kit is as good as Cisco’s, but most of us have never had the chance to ‘play’. This is a real chance for them to put out virtual ‘lab’ editions of their products, and show us how good they are.
The loss of dynamips will be a loss to the engineers who want to learn more about Cisco technology, but it will help them move away from thinking that ‘there is only Cisco’. This will bring about better diversity of skills from engineers, which in turn may hurt Cisco’s dominance in what is after all their core business.