Hurricane Electric's IPv6 Tunnel Broker Forums

Tunnelbroker.net Specific Topics => Questions & Answers => Topic started by: cholzhauer on October 14, 2009, 08:16:29 PM

Title: Uptime
Post by: cholzhauer on October 14, 2009, 08:16:29 PM
I'm using a Sixxs tunnel for IPv6 access. Well, I was.  It's been down all week and I haven't heard anything about when it might come back up. 

I have two questions

1) What is everyone's experience with an HE tunnel?  I'm sure they don't have 100% uptime, but have you experienced any down time?

2) I have to imagine there are a limited number of "backbone providers" (for lack of a better term) that hand out address, routes, ect.  Do HE and Sixxs share the same ones?
Title: Re: Uptime
Post by: comptech on October 14, 2009, 09:36:00 PM
I've been using a few HE tunnels for about 4 months and I would guess the uptime has been around 99%.  I've been told that the issue that caused my tunnel to go down a few times briefly has now been resolved, so it should be even better now.
Title: Re: Uptime
Post by: broquea on October 14, 2009, 10:02:34 PM
Regarding backbones and IPv6 allocations:

HE.NET is the backbone provider in our case. We issue addresses from our /32, and do not use other networks for IPv6 transit or address space. We peer directly with other networks and also provide IPv6 transit. We operate every single tunnel-server and their upstream routers, etc. (again, we maintain our own international dual-stack backbone).

In comparison, I believe Sixxs uses several partner ISPs and organizations to donate bandwidth and hardware resources, along with address space assignments from those organizations (please correct me if I am mistaken). Not that anything is wrong about this, it is simply not how we operate our tunnelbroker service.

********************

Regarding uptime:

While I couldn't say any free service has 100% uptime, we do actively monitor the tunnel-servers and if a problem is detected, network engineers are alerted and work to resolve the issue. Sometimes this might involve relying on a remote facility's personnel (typical remote hands stuff, reboots, etc.).
Title: Re: Uptime
Post by: jimb on October 14, 2009, 11:06:20 PM
FWIW, from my experience HE has been extremely reliable. 
Title: Re: Uptime
Post by: snarked on October 14, 2009, 11:49:06 PM
Having used HE, both via tunnel service and as a native IPv6 carrier (via my co-location provider), I have found no noticable downtime in 20 months of use.

My native IPv6 service is currently down due to equipment failure of my co-lo provider, not HE's fault.
Title: Re: Uptime
Post by: cholzhauer on October 15, 2009, 05:01:46 AM
Quote
In comparison, I believe Sixxs uses several partner ISPs and organizations to donate bandwidth and hardware resources, along with address space assignments from those organizations (please correct me if I am mistaken). Not that anything is wrong about this, it is simply not how we operate our tunnelbroker service.

Correct.  Sixxs uses "pops" around the world that allocate their address ranges and provide transport, that is why I asked about HE.
Title: Re: Uptime
Post by: Ninho on October 15, 2009, 02:01:40 PM
Don't know about SixXS  - OTOH, I have yet to see HE's tunnel fail me  ;D

Outstanding service, wish every other bit of the internet, free or not free, were as reliable as this !





Title: Re: Uptime
Post by: brad on October 15, 2009, 05:09:22 PM
The over all reliability of HE's service as well as customer interaction is like night and day over the SixXs project.

HE provides a top notch well peered network (although not perfect) and redundant network. A single point of contact for support issues which is more or less treated as a production service. Support tends to be very good, no complaints there. The POP servers are very reliable, so is the network. If the tunnel broker POP server dies (hw) they typically replace the system within a day or so.

SixXs consists of POPs provided by partner ISPs. The service varies from POP to POP. The networks are not necessarily very well connected and especially if there is only a single transit provider it can be somewhat unreliable. The people running SixXs can be a real problem to deal with and they have a somewhat strange system for running the tunnel broker, points for doing things, wtf? So support isn't so great and it is not even close to being treated as a production service and so often replies are of the tone, who cares? If a tunnel broker POP server dies (hw) it could be weeks if not never returning, which has happened with some of the U.S. POPs.
Title: Re: Uptime
Post by: cholzhauer on October 15, 2009, 05:51:29 PM
im seeing the same things...my tunnel has been down since monday. 

ill be starting the changeover tomorrow.

one nice thing about sixxs is that they offer support for dynamic tunnels, so if you cant pass protocol 41 for whatever reason, you can make one of those tunnels work pretty easily