Could this little device be the solution to the s@h campus fibre woes?

No new fibre required, likely no new routers or any other equipment required. Easy.


Too good to be true?

See:

D-Link 1000BaseT  to 1000BaseSX Multimode Media Converter  	 	$351

	1000BASE-T Gigabit Twisted-pair to 1000BASE-SX
	Gigabit Fiber Multi-mode Fiber (550m, SC) Media Converter Module 	

• UTP to Multimode Fibre Media Converter
• SC Fibre connector 	• Includes Power Supply for standalone use

DMC-700SC

D-Link 1000BaseT  to 1000BaseLX Singlemode Media Converter  	 	$1,117

	1000BASE-T Gigabit Twisted-pair to 1000BASE-LX
	Gigabit Fiber Single-mode Fiber (10km, SC) Media Converter Module. 	

• UTP to Singlemode Fibre Media Converter
• SC Fibre connector
• Distance up to 10km 	• Includes Power Supply for standalone use

DMC-810SC


One of them at each end gets you 1Gbit/s over 2km of multimode fibre.

That looks to be a drop-in replacement and an awful lot less than $100k!


(s@h has a 1Gbit/s internet link. However, the campus fibre connection to that link only runs at 100Mbit/s at present...)

Happy crunchin',
Martin


Also posted on the Technical forum

---

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

You should PM Matt and see if he responds.

Hi, sounds almost too good to be true, but it's a hell of a lot cheaper, then the 'old' solution.

---

Problem is it's not so simple.

The infrastructure is run by the University IT department and they have their own standards for routers. They will have management systems for monitoring and programing the dozens of routers around the campus. They will not just plug in any random boxes and hope they work.

In a small organisation, where you are THE I.T. manager or consultant, you can just go and buy the boxes you need and set them up. In a larger organisation, life is not so easy.

The Dilbert comic strips pretty much explain how it all works.

Ian

The Dilbert comic strips pretty much explain how it all works.

Dilbert explains pretty much how everything works.

---

The Dilbert comic strips pretty much explain how it all works.

Dilbert explains pretty much how everything works.

Or as It doesn't work as the case may be. :D

---

The T1 Trust, PRR T1 Class 4-4-4-4 #5550, 1 of America's First HST's

Problem is it's not so simple.

The infrastructure is run by the University IT department...

Indeed so.

And hence part of the excitement of this solution is that those little boxes may well just be a "transparent" inline solution for the existing equipment. That is very likely so at the SSL end with their Gbit routers. Also, there have been previous hints that there is a Gbit route possible for the Hurricane end...

The real excitement is the claim made for the multimode fibre adapter!...

Whatever... It's a possible neat low cost solution if it fits into the overall picture.

Happy crunchin',
Martin

---

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

And hence part of the excitement of this solution is that those little boxes may well just be a "transparent" inline solution for the existing equipment.

That's precisely what these are. Simple electical-to-optical conversion boxes. As long as the fibre in place supports the data rate (and most likely it does) it's unlikely Berkeley IT would have a legitimate objection, especially if they were given a chance to vet the installation beforehand.

Go for it, I say. Of course, we'll run into the same old problem -- you throw more disk space or bandwidth at the users, and pretty soon the need grows to fill the capacity.

---

That's precisely what these are. Simple electical-to-optical conversion boxes. As long as the fibre in place supports the data rate (and most likely it does) it's unlikely Berkeley IT would have a legitimate objection, especially if they were given a chance to vet the installation beforehand.

... and the fundamental misunderstanding: IST won't "vet" the installation beforehand -- they'll be the ones specifying the gear and doing the installation.

This is a service that IST provides to SETI@Home, and if this plays out, it could make IST's life easier.

IST seems very user-friendly, but this is their plant, their facilities.

---

... and the fundamental misunderstanding: IST won't "vet" the installation beforehand -- they'll be the ones specifying the gear and doing the installation.

This is a service that IST provides to SETI@Home, and if this plays out, it could make IST's life easier.

IST seems very user-friendly, but this is their plant, their facilities.

This may sound more negative than my intent.

What I'm trying to point out is IST doesn't let just anyone do anything they wish, or their network would be a huge, tangled, unsupportable mess of, well, everything.

The upside of this is that IST might have a bunch of links that would benefit from something like this, and if they called up D-Link with a potential sale of 50 units, they could probably get a couple of them as "evaluation" units, since IST would likely buy a small pile of them and upgrade other links instead of spending a ton of money trenching for fiber.

(Edit: before you get too excited, getting the fiber to work is just part of the problem -- Matt says faster routers are needed too.)

So, we generate buzz here. Someone at SETI@Home adds that "buzz" to the ticket with IST, and from there, we hope it works out.

---

Unless there's an existing single mode fibre, the idea doesn't work anyhow. The multimode DMC-700SC is only good to 550 meters maximum with the best multimode fibre. I'm sure IST knows what they have in place, and if it were merely a matter of converting copper to fibre at 1GBits it would have been done by now.

OTOH, the project probably won't need 1GBits for some years. An option to add another 100 MBits/second link down the hill should be interesting.

                                                              Joe

Unless there's an existing single mode fibre, the idea doesn't work anyhow. The multimode DMC-700SC is only good to 550 meters maximum with the best multimode fibre. I'm sure IST knows what they have in place, and if it were merely a matter of converting copper to fibre at 1GBits it would have been done by now.

OTOH, the project probably won't need 1GBits for some years. An option to add another 100 MBits/second link down the hill should be interesting.

Or a few more increments of 100Mbit/s with 100Mbit/s media converters if the existing fibre cable is actually a fibre bundle...

(Was there ever any comment on whether the fibre is multimode or singlemode?)

Happy crunchin',
Martin

---

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

(Was there ever any comment on whether the fibre is multimode or singlemode?)

I asked that question, specifically of Matt, in the parallel Technical News thread where Martin also posted the D-Link information. But all of this discussion has taken place over Matt's weekend, so we wouldn't have expected a reply yet.

More interestingly, I did a quick google for media converters (they're an absolutely standard bit of kit, if not yet common for domestic installations, and every serious networking manufacturer will have a comprehensive range). One of the first I looked at offered a model - specifically, a GBIC (the interchangable module which actually connects to the optical fibre) - capable of driving multi-mode fibre at 1 Gbit/sec over 2km. That brings us into the realm of the possible, if the other characteristics of the fibre, particularly wavelength, are right - the D-Links are too short-range over multimode, as Joe says.

before you get too excited, getting the fiber to work is just part of the problem -- Matt says faster routers are needed too

Faster routers would be needed to get up to the full 1 Gbit/sec, but we don't need that yet. Remember what happened last time? We had a 100 Mbit/sec link, but it was maxxing out at about 60. After much scratching of heads, somebody eventually read the spec sheet of the routers then in use, and found that the CISCO engine which handles the tunneling was slower than all the other port and routing specifications. They installed new, higher-rated routers and that got us to the 95 Mbit/sec we can use today - this time, we presume, limited by the fibre and line drivers. IIRC, the new routers were specced with a reasonable amount of spare capacity (very good judgement by whoever it was that donated them - he's credited somewhere) - a fibre/driver upgrade should use some of the spare capacity, and then a router upgrade would give the next boost. tick, tock.

(Edit: before you get too excited, getting the fiber to work is just part of the problem -- Matt says faster routers are needed too.)

Not in the lab, they're not. Found the reference in Happy MLK Day (Jan 22 2008): "Bill Woodcock ... has donated another cisco router to us to replace [the] weaker 2811. It a 7600 series, a bit overkill, but will give us tons of headroom to spare."

Looking at the CISCO 7600 spec sheets, the weakest link (like last time) should be the VPN tunneling adapter, which I guess will be a Cisco IPSec VPN SPA - rated at 2.5 Gbps per module. That should keep us going!

NB: Matt didn't say what had been installed by IST at the bottom of the hill to handle their end of the tunnel. That one may not have been upgraded as far.

(Edit: before you get too excited, getting the fiber to work is just part of the problem -- Matt says faster routers are needed too.)

Not in the lab, they're not. Found the reference in Happy MLK Day (Jan 22 2008): "Bill Woodcock ... has donated another cisco router to us to replace [the] weaker 2811. It a 7600 series, a bit overkill, but will give us tons of headroom to spare."

Looking at the CISCO 7600 spec sheets, the weakest link (like last time) should be the VPN tunneling adapter, which I guess will be a Cisco IPSec VPN SPA - rated at 2.5 Gbps per module. That should keep us going!

NB: Matt didn't say what had been installed by IST at the bottom of the hill to handle their end of the tunnel. That one may not have been upgraded as far.

Good sleuthing there!

Looking good!!

Happy crunchin',
Martin

---

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

(Edit: before you get too excited, getting the fiber to work is just part of the problem -- Matt says faster routers are needed too.)

Not in the lab, they're not. Found the reference in Happy MLK Day (Jan 22 2008): "Bill Woodcock ... has donated another cisco router to us to replace [the] weaker 2811. It a 7600 series, a bit overkill, but will give us tons of headroom to spare."

Looking at the CISCO 7600 spec sheets, the weakest link (like last time) should be the VPN tunneling adapter, which I guess will be a Cisco IPSec VPN SPA - rated at 2.5 Gbps per module. That should keep us going!

NB: Matt didn't say what had been installed by IST at the bottom of the hill to handle their end of the tunnel. That one may not have been upgraded as far.

The link from Hurricane Electric to SSL is a SERVICE provided by IST.

What we're talking about are products.

Let me give you an example.

I have a T1. It has a proper RJ48, it looks like a normal B8ZS/ESF wire, with great characteristics. A "real" T1 would be a four wire connection from here to the CO, with about three amplifiers. I connect to it with a CSU/DSU and V.35 serial just like any other T1.

What Verizon actually installed is a jack with some electronics, and a single pair to a matching set of electronics at the Central Office. The wire is actually running HDSL2 at slightly faster than T1 speed. It runs error correction so that most errors get corrected and it's probably slightly better than a "real" T1. It takes less equipment so there is less equipment to fail.

In other words, "what I get" is T1 -- how is "HDSL2."

IST will respond to the open ticket with either "yes" or "no" and a formal cost for the upgrade. What SETI gets is "gigabit" -- how is up to the service provider.

... and when it happens, operating that link and keeping it operational will be up to IST -- it will be their responsibility.

---

Matt has been notified of this thread and will post when he is able I believe :)

---

Honestly, I'm purposefully staying out of this loop as I'm already in way too many loops. But my general understanding of the current situation is:

1. Campus only uses hardware it tests/trusts *and* are willing to use in bulk (so they don't have to deal with a zillion variants).

2. This bottleneck was a major problem a month or two ago, but due to increasing the resolution of the science (and thus slowing down multibeam workunits) our bandwidth needs were vastly reduced.

3. Nevertheless we are still waiting to hear from campus before we can really do anything. At least that's where I think we are.

As always, thanks for your concern/interest/ideas.

- Matt

---

-- BOINC/SETI@home network/web/science/development person
-- "Any idiot can have a good idea. What is hard is to do it." - Jeanne-Claude

Hey... Thanks for taking a look-see.

Can't help but get excited over a possible $4000 transparent fix as opposed to a $100000 bureaucratic saga ;-)


Good luck for the Campus Networking people looking favourably upon the project.

Regards,
Martin

---

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

By the way, given the state budget crisis I'm sure central campus computing is vastly affected by this, and I wouldn't be surprised if *all* such improvement projects are put on hold. We shall see...

- Matt

---

-- BOINC/SETI@home network/web/science/development person
-- "Any idiot can have a good idea. What is hard is to do it." - Jeanne-Claude