If you would like to be listed on this page, send the information to press@openbsd.org.
NOTE: For reasons of security, companies can ask us to withhold their names, or those of their clients. They would then appear as "Undisclosed Company".
AMS-IX deployed two OpenBGPD route servers end of 2009. Each is currently running with over 350 route server client sessions, altogether announcing over 38000 prefixes. The support we got from the developers over the past year has been amazing, every bug or issue found has been fixed as soon as possible and features have been added for us. Thanks and keep up the good work!
Our customers expect us to maintain a high degree of performance, security and control on our systems and network. At the same time we must remain cost effective. The OpenBSD bgpd allows us to do exactly that, and it forms a mission critical component of our business. The skill and dedication we've seen from it's developers is unparalleled.
I first switched two of our routers from Zebra to OpenBGPD after finding a problem we suspected was in Zebra. As it turned out, the problem itself was deep inside OpenBSD, solved in the then-current kernel, but once I'd made the switch I liked what I saw and we've stuck with OpenBGPD.
I've found some bugs, but unlike in some other projects they were resolved rather quickly in spite of being difficult to reproduce, and OpenBGPD seems quite stable now. Thanks, guys!
We have been using OpenBSD for all of our routing, bridging, internet servers, firewalls, VPN, etc. since we started the company. Our growth had forced us to install a second fiber connection and implement BGP. We have been running bgpd on OpenBSD since August 2004, and bgpd has been extremely stable. To date, we have not had any problems — it just works. We run bgpd on a P3 866 with 512 MB of RAM. We route over 800 GB of internet bound traffic every month. We are very impressed with the work that the OpenBSD project does, and we appreciate the quality, stability, functionality and security that the operating system provides for us, but more important, we are grateful and appreciate the hard work and dedication of the team. Thank you OpenBSD.
The first thing I said was let's use Junipers, but after seeing bgpd in action on a test machine I was hooked.
We were in the situation where we needed to upgrade from a Cisco 3600 due to the expanding bgp routing table and a seriously lacking cpu. To do what we needed to do would have required a Juniper or a high end Cisco. Instead we tried OpenBSD with OpenBGPD and pf on a low end Opteron processor. The result was nothing short of shocking! Three full views with 700 Meg of memory to spare. CPU utilization between 5% and 20%. The ability to alter traffic filters using pf without worrying about the CPU. Full Unix-like environment with things like tcpdump. In short, it's the best hardware, software decision we ever made and it has literally changed the services we can offer. I very much prefer OpenBSD, OpenBGPD, and pf to any other solution. It actually brought tears to my eyes during our first few days of use. It's that good.
Why tax a router with gathering statistical or operational data when OpenBGPD can do it natively quickly and efficiently? OpenBGPD fulfills the OpenBSD promise of providing reliability and security over flashy buttons and knobs. I look with great interest at this project and have high hopes for its continued success.
I just wanted to tell that I am a very happy user of OpenBGPD (and OpenBSD for years). We started out with zebra on OpenBSD 2.8 or 2.9 and OpenBGPD since 3.8. Now we run OpenBSD 3.9 (since yesterday) with OpenBGPD. Everything even remotely important runs OpenBSD here ;) We also provide businesses with dsl and soekris/openbsd routers etc etc. We are moving to a new server room at one of the facilities that host the AMS-IX (and NL-IX) exchanges and I will set up redundant CARP/BGP (maybe OSPF) for our network there. We have around 160 peers and three transit providers and my life is sooo much easier than it was with zebra! Thanks dude!
We just set up a local AS using OpenBSD/flashdist/OpenBGPd and the experience was really great, simple "plug and play". The remote cisco & juniper stuff never complained and everything has been working flawlessly since then. Thumbs up to the OpenBSD / OpenBGPd people !
As a 'full service offering' company, every aspect of our business is critical. Hence using Cisco routers was a simple choice 5 years ago when we started offering enterprise hosting. But as our clients needs continue to grow, our service offerings also need to improve. Looking at upgrading our network infrastructure recently, I was shocked at the cost of implementing BGP on our Cisco gear — the extra router, the RAM upgrades, the maintenance! Having used OpenBSD as our internal firewall for over 7 years, I was familiar with it, so decided to give OpenBGPd a try and WOW, its impressive. Replacing our core router was SIMPLE, and it just worked! For the cost of just the Cisco RAM upgrade and one year support, we have a complete OpenBSD / OpenBGP based BGP infrastructure.
We mainly provide web-, image- and media- services for other designers and publishers. We were unhappy with our old Cisco router that was ineffective and biased traffic incorrectly over transit lines. It didn't support a full BGP table and crashed if we added too many peers. Now we run OpenBGP, with full tables and peering with approx 40 peers (at NIX), and traffic runs much smoother.
We are currently using OpenBGPD in a High Availability Routing Solution in our backbone. We have changed from Cisco to Linux+Zebra and now to OpenBGP and we are very satisfied.
We handle a "small" academic network whose connectivity relies on about 150mbit of bandwith carried over 3 BGP peerings feeding us the full routing table. We serve a cluster of cancer research institutions which live mostly with public donations and research grants.
The "routers" for this are two entry-level Sun/amd64 boxes costed less than 1000 euro each which run FreeBSD and carp/OpenBGPD. We never saw the load on the machines and/or memory grow beyond 5% on the main router (the other just sits there waiting).
We are incredibly happy of the performance, stability and simplicity of OpenBGPD and we really thank you for this piece of software which saved us tens of thousands of euros that have been thus used for research activities and otherwise would have gone in the pocket of some big company now. The only time we had a configuration problem we wrote to the OpenBSD mailing list and got an answer with the solution 38 minutes later, that "big company" support team should learn something.
In every city we have 2 OpenBGPD route-servers with fullytransparent services — local pl-ix (peering between companies in one city) and global pl-ix (peering with every company in PL-IX). BGP sessions are made between member and PL-IX route-servers.
Our transit router (transits traffic from/to other Polish IX - WIX), also based on OpenBGPD, serves 300Mbps in + 300Mbps out traffic with ~100kpackets/s in and ~100kpackets/s and full bgp sessions, and since we started it we forgot about that box — it just works.
OpenBGPD has given us very stable and reliable platform for PL-IX members. During PL-IX startup help of OpenBGPD developers was extremely useful, and what is the most important, response times for our problems were amazing. Big companies with well-payed-support should learn from you.
From the start we relied on OpenBPGD. It has proven itself as very stable, no incidents whatsoever. Big compliments to Mr. Brauer!
We have been using OpenBGP since shortly before OpenBSD 3.9 was released. We were upgrading our Internet connections and our existing Cisco router was already struggling. We were looking at having to buy a big new Cisco router to get the performance we needed. We were instead able to buy three Dell servers and run OpenBGP on them for much less cost and better redundancy than the Cisco router. The OpenBSD solution has better performance, and as we continue to add bandwidth, we should be able to continue using the current servers. We are running three full views and everything Just Works.
Our entire BGP setup has been running FreeBSD/Zebra for a number of years until we made the transition to OpenBGPD approximately 2 years ago [Jan. 2005]. Besides an immediate performance leap we gained significant production advantages with tools like 'bgpctl' and 'pfctl' in our multihomed environment. OpenBSD has with it's OpenBGPD, OpenOSPFD and OpenNTPD provided us with a rock-solid and truly "One Stop Shopping" BGP solution for our ever growing business-to-business hosting company.
We are a managed service provider, delivering both front-office SaaS and traditional hosting services across the EMEA region. High availability is a big deal for us and our clients. As a result, in the event of any downtime, our SLAs are backed off with compensation in hard cash directly from our revenue stream.
In the first quarter of 2005 we took the decision to become an LIR and provision our own multi-homing. Having already been familiar with the reliability and performance of PF/CARP, as well as wishing to avoid the overhead and vendor lock-in of certain big names, we gave OpenBGPD a go.
Three years later we are taking a number of full feeds, have seen out provider outages and rolling upgrades, all without breaking a sweat or disruption to our services. Many thanks to the OpenBSD team.
Aquilenet is a non profit organisation and a "do it yourself ISP", member of a Federation of similar ISP in France called FFDN. We are netneutrality builders, helping for more freedom and building networks.
Since 2011 our backbone network is stable due to our OpenBSD build on CARP/pfsync/OSPF/pf magic, and our transit based on OpenBGPD. The work of the community is quite impressive and we are really happy to contribute to these projects. We encourage all ISP to use OpenBSD/OpenBGPD as simple as efficient !