Congratulations! Well done! I´m surprised how small the disturbance was for that huge change!

Jeff and Court deserve about 98% of the credit. I purposefully stayed out of the loop until after Court left and then I was forced into the loop.

mdadm means software-raid, right? Is it the money-problem why you don´t use a raid-controller or are there other reasons?

We've been bitten by hardware raid in the past (causing random database corruption in one case, not effectively reporting failures/errors in several others), so I'm a bit biased. In my experience, I feel the control and access of software raid far outweighs any performance loss from not using hardware. I'm the kind of guy that hates hidden levels of anything - I like to know exactly what's going on how and when and where and why at all times, and you don't seem to get any of that from a hardware raid device. This is also why I despise object oriented programming. I grew up having to peek and poke Pet and Apple computers to get anything to work, so I can't stand having these details hidden from me. Hiding important details in hardware and black-box software defies true understanding, and therefore turns computing from an art into, well, something that isn't very different from bureaucracy. Wow. That was a mini-rant. A had a couple sips of coffee this morning. Sorry.

- Matt

I would have to say that I agree with you here. I don't do much computing work know, but I did quite a bit in my younger years .. in fact my first computer was a TI-99 when I was 11, anyone remember those .. I learned all of my programing there and on several first run apples before the mac came out. Although I love the ease of use with computers now, there is alot going on behind the seans that you dont see. In the early years it was much different. Well love the work you guys do .. keep it up .. I want to find an ex-t because maybe he will take me with him .. lol ..

---

Hurricane Electric (HE) offered us 10 times the bandwidth at one fourth the price

So, does this mean you're now running gig-E on setiboincdata? Or are we still limited to Fast Ethernet? Looking at the Cricket graphs, I see a peak of about 60mpbs even during the heavy recovery from the downtime, and I was getting some dropped connections at that time. What is the bottleneck?

---

...I see a peak of about 60mpbs even during the heavy recovery from the downtime, and I was getting some dropped connections at that time. What is the bottleneck?

In the past, the bottleneck has been the number of simultaneous connections that the server can keep open (due to memory resource limitations?) and the speed at which the database server could service requests.

I would expect that the (solaris?) simultaneous connections limit is still there...

(Or is there a NAT box in the mix that runs out of table space?...)

Happy crunchin',
Martin

---

See new freedom: Mageia Linux
Take a look for yourself: Linux Format
The Future is what We all make IT (GPLv3)

Well, it was just posted by Matt that bruno is now the upload/download server (I assume that means it is setiboincdata.ssl.berkeley.edu--tracerouting that does go over the HE link). So, hopefully, Gigabit Ethernet + 2 x 2.8GHz Xeon and 12 GB RAM = a better recovery rate after outages. (And looking at network graphs peaking at multi-hundred mbps speeds is always fun!)

I'm pretty sure NAT is waay beneath these guys...

---

Nope, Not according to this link, Kryten still is, If that's a mistake somebody ought to fix It. Bruno is the file deleter now there and that looks like progress is being made on the Kryten problem. :D

I was looking at this post from yesterday, in which Matt says bruno is now the upload server. I noticed that the server status page still said kryten was, but I'm assuming that just hasn't been updated yet.

Edit: Matt says this in that post:

Fair enough - we'll leave both kryten and bruno up as "mirror" servers as DNS (hopefully) corrects itself over the coming days. I'll reflect the changes in the server status page eventually.

So with two upload/download servers, we should really be able to see some great thoroughput! :-)

---

> see Matt's NEW POST - Ups and Downs (Mar 22 2007)

Time passed, and with inflation this deal became less and less affordable. Eventually we had to start looking elsewhere. Hurricane Electric (HE) offered us 10 times the bandwidth at one fourth the price, so we started moving in this direction. This was about 18 months ago.
- Matt

The Cogent pipe should have been 100 megs for $1,000 (for wholesale users, its $1500). So you're telling me that HE gave you 1 gig for $250 - $375?

---

Jeekz, that is expensive...
I have 100megs and pay just 480$ per year with free traffic.

Carl

Jeekz, that is expensive...
I have 100megs and pay just 480$ per year with free traffic.

Carl

There's a big difference between broadband and dedicated bandwidth. Cogent and Hurricane are among the least expensive dedicated bandwidth providers.

I bet if you used your whole 100 megs for any extended amount of time, your ISP would have some words with you.

---

Phewww! ,, Good job guys!!

Thought for a minute there my 2 line, multilinked, static IP, serverd thru ICS, DIAL-UP, was draining the whole SETI network down.

JK'ing, Glad to see progress still being made for the SETI team.

Congrats!

TimBo

Jeekz, that is expensive...
I have 100megs and pay just 480$ per year with free traffic.

Carl

There's a big difference between broadband and dedicated bandwidth. Cogent and Hurricane are among the least expensive dedicated bandwidth providers.

I bet if you used your whole 100 megs for any extended amount of time, your ISP would have some words with you.

Actually they haven't so far. That is the actual cost we are paying at the company where I work.

Jeekz, that is expensive...
I have 100megs and pay just 480$ per year with free traffic.

Carl

Unless your ISP is able to maintain peering relations with most major networks, it is costing them a bunch for that connectivity. Broadband networks are greatly oversold (some times in excess of 100:1), so that is where they are often able to sell you a cheaper service.

Many of the more expensive networks could be 3x the cost of Cogent for a comparable service.

---

Actually that is also depending on which country you are in and the amount of fibre available per capita.
I can get it as a broadband (oversold) at about one tenth if I shop around. 100MB dedicated lines is though considered pretty small nowadays so they are cheap. A one GB would be more expensive per MB...

Actually that is also depending on which country you are in and the amount of fibre available per capita.
I can get it as a broadband (oversold) at about one tenth if I shop around. 100MB dedicated lines is though considered pretty small nowadays so they are cheap. A one GB would be more expensive per MB...

The country makes more of a difference on retail than on wholesale accounts. Many carriers charge the same rate whether you buy it in the US, UK, or Japan. Now on retail, yes, that varies widely, though it is more technology based than country based. FTTx in the US costs about the same as it does in Sweden or anywhere else, it just isn't as widely available.

$/mbit also is a difference in wholesale and retail. It almost always goes down the more you buy in wholesale. However, retail companies know there is less competition so can charge you an increasing rate taking advantage of you.

BTW: My resume on this is that I am a US based ISP and have sourced wholesale bandwidth all over the world. I also speak with other ISPs throughout the world.

---

Well we are still pushing 10mb on average, peaking around 40 and still no complaints from our ISP.

My resumé, obviously good at making deals;-)

Could those posting in this thread clarify whether they are talking about symmetric download/upload links, please? The orginal discussion was with regard to the SETI labs, where the vast bulk of the bandwidth is used for outbound data. Anyone making comparisons with consumer services (esp. DSL) should remember that these are optimised for inbound traffic.

Well we are still pushing 10mb on average, peaking around 40 and still no complaints from our ISP.

My resumé, obviously good at making deals;-)

*sigh* Your resume is that you're John Doe consumer that doesn't know the Internet world beyond retail consumer services.

---

Could those posting in this thread clarify whether they are talking about symmetric download/upload links, please? The orginal discussion was with regard to the SETI labs, where the vast bulk of the bandwidth is used for outbound data. Anyone making comparisons with consumer services (esp. DSL) should remember that these are optimised for inbound traffic.

I am talking about wholesale, dedicated fully symmetric bandwidth. The other poster is referencing a retial, consumer service. It may be symmetric, but it cannot be compared to a dedicated service.

Wholesale service:
Service performance (speed, loss, jitter, latency) is guaranteed by an SLA
Unrestricted use of the service
BGP available
IPs provisioned in real subnets, and thus routable
Any speed one wants is available
Symmetric

Broadband service:
No or a very limited guarantee
Restricted use (Usually can't use many kinds of servers, fair use policies, service throttling)
No or very limited external routing
IPs provisioned are in generic blocks, not able to be subnetted or routed
Only speed available is speed advertised
Usually asymmetric

---

Well we are still pushing 10mb on average, peaking around 40 and still no complaints from our ISP.

My resumé, obviously good at making deals;-)

*sigh* Your resume is that you're John Doe consumer that doesn't know the Internet world beyond retail consumer services.

Yes, my resumé is that I am a John Doe consumer who checks around with every available ISP provider with a very precise spec, then choosing the cheapest serious provider.
Please do not make the mistake of assuming that a retail customer is uninformed. That might be true in many cases, but not in every case. You will only loose customers if you do.
FYI, we use OPTOSunet as ISP.

Could those posting in this thread clarify whether they are talking about symmetric download/upload links, please? The orginal discussion was with regard to the SETI labs, where the vast bulk of the bandwidth is used for outbound data. Anyone making comparisons with consumer services (esp. DSL) should remember that these are optimised for inbound traffic.

I am talking about wholesale, dedicated fully symmetric bandwidth. The other poster is referencing a retial, consumer service. It may be symmetric, but it cannot be compared to a dedicated service.

Wholesale service:
Service performance (speed, loss, jitter, latency) is guaranteed by an SLA
Unrestricted use of the service
BGP available
IPs provisioned in real subnets, and thus routable
Any speed one wants is available
Symmetric

Broadband service:
No or a very limited guarantee
Restricted use (Usually can't use many kinds of servers, fair use policies, service throttling)
No or very limited external routing
IPs provisioned are in generic blocks, not able to be subnetted or routed
Only speed available is speed advertised
Usually asymmetric

1. Our service is guaranted in 100MB symmetric.
2. unrestricted
3. Subnet routable.
4. Not every possible speed available. For the moment there is an upper expansion limit at 10GB.

Anything else you need to know?