convert to existing fiber uplink from existing copper

[Normal Mouse]

We have an old Cisco 2620 router with what we think is a 1Gbps copper RJ45 line coming in from "the backbone". We're part of a large organization, and we don't know exactly what's on the other side of our lines. There is much infighting, and asking questions might lead to political turmoil we don't need. So, I apologize in advance for our ignorance.

This is obviously an old and slow router, and recently traffic is causing gluts that manifest as intermittent outages. They are starting to be longer and more noticeable, so we're looking to upgrade.

In an ideal world, we'd just just get new hardware and things would get 10 times faster and we'd only run into this problem again years from now, but that's not a reality.

We have noticed that there is a fiber line, which we believe is running directly back to the mysterious "backbone" and we think is active. Presumably, this is faster than the copper.

Previously we had a 100Mbps switch in between our Cisco and the copper "backbone" line coming out of the wall. It shouldn't have mattered much, since our Cisco can only handle 100Mbps, but switching to a cheap 1G switch magically made our network much faster. We are wondering if the same magic could happen if we could somehow plug the Cisco into the fiber "backbone" line.

Is this theory even slightly plausible? (Keep in mind that we don't know who we're sharing the copper line with on the other side.)

Assuming for the moment that it is, what would be necessary to connect the Cisco into the fiber outlet in the wall?

We have seen converters online for about 150 bucks, but the purpose of these seems to be to convert from copper to fiber, through a fiber extension, then from fiber back to copper. The implication is that you need two of these things to make them useful. So, is it a fiber-to-copper modem that we need?

Those are significantly more expensive. Note that we don't really know if the fiber line is live, so we want to keep testing expenses to a minimum. Is there a cheap way to test the line before we dump (what is for us) significant money on the problem?

Sorry we don't have better information. Thanks very much for your time.

 

[Ralston18]

However you end up connected/connecting the network will only be as fast as the slowest link in the path. Copper between fibers is going to be copper speed.

Despite the infighting and politics you need to find a management champion who can understand the problems and make a case for the desired fiber connectivity. Basic ROI, cost/benefit approach.

My reason for that is, if you do not have everything properly approved and agreed to beforehand, the fiber might later be deactivated. Accidently, maliciously, for security reasons...whatever. Then you are out of the connectivity, the money, and the time spent. Not a good place for any manager.

In fact you may end up in trouble if you make some unauthorized connection. Security is, and should be, taken very seriously these days. Violating company rules, policies, etc. can even lead to firing and/or legal action.

There are fiber testers out there but you need access to both ends especially if the fiber was installed as "dark". Meaning installed for future use but never hooked up.

It is harder way but do things properly and aboveboard. Hopefully there are some reasonable folks around.

 

[Normal Mouse]

Hopefully there are some reasonable folks around.

Of course your advice is solid, but if there were reasonable folks around, we wouldn't be posting anonymously, hoping for tech help from strangers!

We appreciate your post, but we'd still like to find out what our options are from a technical standpoint.

 

[bill001g]

You are going to have to guess blindly unless someone knows where the fiber connection goes. It does not even have to be ethernet if it is going outside the building to a carrier network. It could be something like a OC48 or other OC connections. This is completely different than gige fiber connections. In addition you have be sure you know the frequency of the light being used. This is based on the type of fiber being used and the type of adapters in the equipment.

I suspect you will have no other option than to try to find someone who know what the connection actually is. You can spend lots of money guessing.

 

[Normal Mouse]

Well, they ran the fiber line to server room years ago with the intent of it being used on what I can only imagine was run-of-the-mill hardware. It's officially our line and we're allowed to use it, but we never got the associated hardware at the time and it's politically impossible to get the proper hardware now.

But let's just say it's packing just the right type of signal. Is there a reasonably priced piece of hardware that could convert that incoming line to a regular RJ45-style copper line that we could plug into an old router?

If so, is that even likely to improve the speed?

Just trying to determine if this line of thought is a total waste of time.

 

[kanewolf]

Just because a network segment is fiber, that doesn't mean it is any faster than copper. Fiber can be 10Mbit, 100Mbit, 1Gbit or 10Gbit. You have to know what the speed of the fiber is to get the appropriate adapter. What you are looking for is a media converter. But if you don't know what the switch on the other end has for a port it would be difficult to recommend a media converter.

 

[Normal Mouse]

OK, so a relatively inexpensive media converter would be all that's necessary, not the modem?

This is a line that we'd expect to plug straight into a fiber-capable router, so whatever's on the other side should be considered in that light. For what it's worth, we think it's 10Gbps.

Maybe the simplest version of our question is, assuming we get lucky with every aspect of what's on the other side of the cable, and let's say it is actually 10Gbps, is it within the realm of possibility that it will make the Cisco 2620 go faster? Well, OK, the Cisco isn't going to go any faster, but is it possible the net result could be a less congested network?

If, even in the best case scenario, that's hardly likely, then it's not worth pursuing. If there's a reasonable chance of improvement, then it's worth experimenting.

 

[Ralston18]

You might get some idea about the situation (i.e., per bill001g - "type of fiber being used") if you can read the specs on the fiber cable end that you can see. Hopefully you can access a long enough length to get enough information.

Could narrow the options down a bit....

 

[kanewolf]

A 2620 is a 100Mbit router. Having a media converter connected to a core switch might improve your performance. BUT, you really need somebody in the network or infrastructure team to assist you. You have no idea what, if anything, is connected to the other end of the fiber. Even if it is connected to something, best practices for networking would have the port disabled.

You may not like playing politics, but it is what will help you most.

 

[Normal Mouse]

We only have a port in the wall, not an actual cable, but I will check out the port and see if there are any clues.

So, OK, it's possible that performance could improve. It's not a wildly stupid thought that it might. That's enough to move forward with research, thanks.

It's not so much an issue of whether or not we like to play politics. The politics have already been played and we were told we're nerds and we should eff off. No more discussion. Those are the cards we've been dealt.

Pushing the issue could lead to someone getting fired, and we don't care that much. We'd like to provide a better experience for the users if we can figure something out, but we're not going to be forced to change our whole lives and maybe move to another city in pursuit of that goal.

The people we work with within our area are reasonable operators we like dealing with, but we're a cog in a giant evil machine and the only way to survive is to keep your head down.

Thanks for your help.