Re: [j-nsp] Juniper vs Arista
On 14/Aug/18 12:22, Gert Doering wrote: > Which is sort of what stuck after our Arista POC: it's like IOS with > lots of useful enhancements, and without the annoying "I refuse to > do things, because I am right and you are wrong" attitude of JunOS. The t-shirt is going to print :-)... Mark. signature.asc Description: OpenPGP digital signature ___ juniper-nsp mailing list juniper-nsp@puck.nether.net https://puck.nether.net/mailman/listinfo/juniper-nsp
Re: [j-nsp] Juniper vs Arista
Hi,
On Tue, Aug 14, 2018 at 10:37:31AM +0200, Sebastian Wiesinger wrote:
> Do they have commit capability?
Yes, optional. You either do "conf term" just like on IOS, or you do
"conf session" which needs commit.
Which is really really nice - if I do troubleshooting, I want a
"interface foo / shut / no shut" to have effect right away without
minute-long commit waits - and for larger change blocks ("bgp neighbor"
comes to mind), commit is good :-)
Which is sort of what stuck after our Arista POC: it's like IOS with
lots of useful enhancements, and without the annoying "I refuse to
do things, because I am right and you are wrong" attitude of JunOS.
gert
--
"If was one thing all people took for granted, was conviction that if you
feed honest figures into a computer, honest figures come out. Never doubted
it myself till I met a computer with a sense of humor."
Robert A. Heinlein, The Moon is a Harsh Mistress
Gert Doering - Munich, Germany g...@greenie.muc.de
signature.asc
Description: PGP signature
___
juniper-nsp mailing list juniper-nsp@puck.nether.net
https://puck.nether.net/mailman/listinfo/juniper-nsp
Re: [j-nsp] Juniper vs Arista
> Of Saku Ytti > Sent: Tuesday, August 14, 2018 9:31 AM > > Perhaps in future it's all very small (<5 0M) net install image, then DEB/RPM > all packages via local proxy to have functioning install for specific target, with > specific forwarding engines. > Cisco is certainly on the right track with the flexible packaging in 64 bit XR (though the mini ISO is nowhere near 5Meg ). adam netconsultings.com ::carrier-class solutions for the telecommunications industry:: ___ juniper-nsp mailing list juniper-nsp@puck.nether.net https://puck.nether.net/mailman/listinfo/juniper-nsp
Re: [j-nsp] Juniper vs Arista
> > > But if I'm honest, I don't put a lot of stock in this. If you remember, > this was Juniper's line back in the day, "One Junos to rule them all". > Well, things obviously took another turn when they introduced the MX and > EX platforms around 2007/2008. > Moreover, as we've discussed a few weeks ago, the "single software über alles" strategy might be just a dumb idea. ScreenOS to JUNOS migration is a good example. Juniper basically lost the firewall as well as small software routers markets by trying to fit all into a single package. And now, look, MX150 is the new J4300 :) Not sure, but from the first glance it doesn't look like they've gained more than they've lost with the JunosE to JUNOS BNG migration. The whole Juniper's enterprise market failure is all about "single JUNOS doesn't fit all": WX, SSL VPN, IPS, Wi-Fi, etc etc etc. And yes, Arista will face the same challenge when they buy someone like Fortinet to expand beyond the niche where they are by the moment. -- Pavel > ___ juniper-nsp mailing list juniper-nsp@puck.nether.net https://puck.nether.net/mailman/listinfo/juniper-nsp
Re: [j-nsp] Juniper vs Arista
On 14.08.2018 08:34, Mark Tinka wrote: On 14/Aug/18 08:26, Gert Doering wrote: Mailman claims "there isn't yet", but if Jared would add one, I'd subscribe :-) https://puck.nether.net/mailman/listinfo/ As would I... +1 -- regards Tobias ___ juniper-nsp mailing list juniper-nsp@puck.nether.net https://puck.nether.net/mailman/listinfo/juniper-nsp
Re: [j-nsp] Juniper vs Arista
On Tue, 14 Aug 2018 at 11:38, Sebastian Wiesinger wrote: > Do they have commit capability? Yes, and it works. In IOS-XR commit doesn't inherently work, because of how it's designed. In JunOS commit inherently works. In IOS-XR BGP team owns commit for BGP config, tunnel team owns commit for GRE, it's not generic infrastructure issue if A=>B doesn't work, that's problem for specific component team. So you can find these impossible configs in IOS-XR quite often, where you cannot move from arbitrary A to arbitrary B. Good acid test for sanity of design 'can I get output of _every_ command in serialised format', <100% coverage means poor and expensive to maintain design. Note, this is not trash on IOS-XR, you can find another argument where JunOS and EOS look poor and IOS-XR looks great, it's merely topical and important-to-me feature to get right. If you're CLI jockey network, like most networks are, then it's not very important thing to get right. -- ++ytti ___ juniper-nsp mailing list juniper-nsp@puck.nether.net https://puck.nether.net/mailman/listinfo/juniper-nsp
Re: [j-nsp] Juniper vs Arista
On 14/Aug/18 10:37, Sebastian Wiesinger wrote: > Do they have commit capability? https://eos.arista.com/config-sessions-tips/ Mark. ___ juniper-nsp mailing list juniper-nsp@puck.nether.net https://puck.nether.net/mailman/listinfo/juniper-nsp
Re: [j-nsp] Juniper vs Arista
On 14/Aug/18 10:30, Saku Ytti wrote: > I know what you mean, but 'code is Cisco-like' may not be read as > positive by every recipient. I think you more accurately mean 'CLI is > like classic IOS'. This... as in, "it's not much to learn". Very different under the hood, obviously, but on top, it's not as foreign as coming from a Cisco world into a Juniper one. > > ANET does not have much run way until they are forced to two image > future, due to 64b control-planes. Not that I care about at all how > many images vendors ship. Lot of it is just smoke and mirrors, like > inflating single package with mostly useless content to make it seem > like single image fits all. Agreed. > Perhaps in future it's all very small (<5 0M) net install image, then > DEB/RPM all packages via local proxy to have functioning install for > specific target, with specific forwarding engines. Probably. Mark. ___ juniper-nsp mailing list juniper-nsp@puck.nether.net https://puck.nether.net/mailman/listinfo/juniper-nsp
Re: [j-nsp] Juniper vs Arista
* Saku Ytti [2018-08-14 10:31]: > On Tue, 14 Aug 2018 at 09:35, Mark Tinka wrote: > > > The Arista code is so Cisco-like, and in our use-case, has done > > everything as advertised in the manuals and in the "Arista Warrior" > > O'Reilly publication. So we haven't had to struggle... but then again, > > our deployment is very simple - Layer 2 core switching. > > I know what you mean, but 'code is Cisco-like' may not be read as > positive by every recipient. I think you more accurately mean 'CLI is > like classic IOS'. Do they have commit capability? Regards Sebastian -- GPG Key: 0x93A0B9CE (F4F6 B1A3 866B 26E9 450A 9D82 58A2 D94A 93A0 B9CE) 'Are you Death?' ... IT'S THE SCYTHE, ISN'T IT? PEOPLE ALWAYS NOTICE THE SCYTHE. -- Terry Pratchett, The Fifth Elephant ___ juniper-nsp mailing list juniper-nsp@puck.nether.net https://puck.nether.net/mailman/listinfo/juniper-nsp
Re: [j-nsp] Juniper vs Arista
On Tue, 14 Aug 2018 at 09:35, Mark Tinka wrote: > The Arista code is so Cisco-like, and in our use-case, has done > everything as advertised in the manuals and in the "Arista Warrior" > O'Reilly publication. So we haven't had to struggle... but then again, > our deployment is very simple - Layer 2 core switching. I know what you mean, but 'code is Cisco-like' may not be read as positive by every recipient. I think you more accurately mean 'CLI is like classic IOS'. ANET does not have much run way until they are forced to two image future, due to 64b control-planes. Not that I care about at all how many images vendors ship. Lot of it is just smoke and mirrors, like inflating single package with mostly useless content to make it seem like single image fits all. Perhaps in future it's all very small (<5 0M) net install image, then DEB/RPM all packages via local proxy to have functioning install for specific target, with specific forwarding engines. -- ++ytti ___ juniper-nsp mailing list juniper-nsp@puck.nether.net https://puck.nether.net/mailman/listinfo/juniper-nsp
Re: [j-nsp] Juniper vs Arista
On 14/Aug/18 08:26, Gert Doering wrote: > Mailman claims "there isn't yet", but if Jared would add one, I'd > subscribe :-) > > https://puck.nether.net/mailman/listinfo/ As would I... The Arista code is so Cisco-like, and in our use-case, has done everything as advertised in the manuals and in the "Arista Warrior" O'Reilly publication. So we haven't had to struggle... but then again, our deployment is very simple - Layer 2 core switching. But I'd certainly like yet-another-place to b**ch and moan about my new vendor, Arista :-)... Mark. signature.asc Description: OpenPGP digital signature ___ juniper-nsp mailing list juniper-nsp@puck.nether.net https://puck.nether.net/mailman/listinfo/juniper-nsp
Re: [j-nsp] Juniper vs Arista
Hi, On Tue, Aug 14, 2018 at 08:20:59AM +0200, Mark Tinka wrote: > On 14/Aug/18 03:41, Jared Mauch wrote: > > > I believe there???s also arista-nsp on here.. > I checked last year and couldn't find anything. Do you have a link? Mailman claims "there isn't yet", but if Jared would add one, I'd subscribe :-) https://puck.nether.net/mailman/listinfo/ gert -- "If was one thing all people took for granted, was conviction that if you feed honest figures into a computer, honest figures come out. Never doubted it myself till I met a computer with a sense of humor." Robert A. Heinlein, The Moon is a Harsh Mistress Gert Doering - Munich, Germany g...@greenie.muc.de signature.asc Description: PGP signature ___ juniper-nsp mailing list juniper-nsp@puck.nether.net https://puck.nether.net/mailman/listinfo/juniper-nsp
Re: [j-nsp] Juniper vs Arista
On 14/Aug/18 03:41, Jared Mauch wrote: > Personally I really like Arista but also like JunOS based on my use case. What we can be certain of is that the small companies become big successful ones, grow, and then have to sacrifice a bit of their earlier principles to maintain the new lifestyle. So it would not surprise me if Arista decide to fork the code depending on what happens with their product lines in the future. To be honest, I've learned that it's not worth trying to tie them down to this; they are running a business after all. Your best defence is to understand the platform, and understand the vendor's behaviours. > I believe there’s also arista-nsp on here.. Is there? I checked last year and couldn't find anything. Do you have a link? Mark. ___ juniper-nsp mailing list juniper-nsp@puck.nether.net https://puck.nether.net/mailman/listinfo/juniper-nsp
Re: [j-nsp] Juniper vs Arista
On 14/Aug/18 03:08, Colton Conor wrote: > We have used Juniper for a long time across various lines. What I have > noticed is on the JTAC recommend version website piratically all of their > product lines are running on different code trains. They all run Junos yes, > but they all have different features. It seems like Juniper has a different > software development team for every product. > > Arista's sales engineering are claiming that their entire product line runs > EOS, and there is a single firmware file for all products. Does anyone know > if that is true, and how that compares to the way Juniper does things? It is true that Arista currently have a single image for EOS across their platforms. But if I'm honest, I don't put a lot of stock in this. If you remember, this was Juniper's line back in the day, "One Junos to rule them all". Well, things obviously took another turn when they introduced the MX and EX platforms around 2007/2008. Ultimately, know your platform, understand your vendor, and don't get caught in the hype. We already have far less time in the day as it is. Mark. ___ juniper-nsp mailing list juniper-nsp@puck.nether.net https://puck.nether.net/mailman/listinfo/juniper-nsp
Re: [j-nsp] Juniper vs Arista
Hi, On Mon, Aug 13, 2018 at 08:08:29PM -0500, Colton Conor wrote: > Arista's sales engineering are claiming that their entire product line runs > EOS, and there is a single firmware file for all products. Does anyone know > if that is true, and how that compares to the way Juniper does things? This is true, so for "control plane" features, you can basically rely on the same things being available everywhere. Now, "data plane" - like, "can this box do MPLS? can it do SVI counters?" - depends a lot on the unterlying hardware. So even if you have the same EOS on top, one box will do MPLS, and another box will not. (And part of that is "product management", that is, the Trident II+ boxes will not do MPLS because "buy a bigger one if you want that", even if the hardware could support some basic set of MPLS features) So: you still need to test box vs. features you need. gert -- "If was one thing all people took for granted, was conviction that if you feed honest figures into a computer, honest figures come out. Never doubted it myself till I met a computer with a sense of humor." Robert A. Heinlein, The Moon is a Harsh Mistress Gert Doering - Munich, Germany g...@greenie.muc.de signature.asc Description: PGP signature ___ juniper-nsp mailing list juniper-nsp@puck.nether.net https://puck.nether.net/mailman/listinfo/juniper-nsp
