To your question, you don't need a switch, and you don't need fiber.  The signal is the signal, so that doesn't change.  The starting point is to connect your Ethernet cable to your streamer and listen to music. From there, if you wish, you can explore the many products designed to "enhance" your digital music experience.  These products generally claim to either remove noise alleged to ride on the Ethernet cables, or to reclock the digital signal thus sending a signal with lower jitter to your DAC.  These range from different cables to switches, to filters, to reclockers, to digital-digital converters (DDC) that buffer and reclock the signal and also allow different formats of digital inputs and outputs.  

Switches can have clocks, such as the English Electric 8 and Bonn Silent Angel N8 switches discussed in this thread, which are both essentially the same switch from the small OEM manufacturer ThunderData, in Guangdong, China company.  They both have the same board inside and both have an added TCXO clock.  My understanding is that the English Electric switch has a fancier case and feet.  There are even more expensive switches you can purchase with one of the top-of-the-line switches being the M12 SWITCH GOLD, by JCAT, for €3,800.  Silent Angel has a Bonn N8 Pro switch, which is supposedly better for audio (BTW, the new fad is to put "PRO" behind a product's name to indicate an upgraded, better quality, and more expensive model).  Here is an article about switches from two years ago.

Regarding fiber, that is simply a method of providing optical isolation for the purpose of reducing noise.  As stated in this thread, you can do that with two $25 converters and inexpensive fiber cable.  However, be aware you are adding two noisy SMPSs with those converters, unless you also splurge for a couple of LPSs at around $100 each.  You can also do it by using the GigaFOILv4, which would require only one LPS.  I have both the fiber/converter set-up, and the GigaFoilv4, and I have tried both, including the LPSs.  My run is 45 feet from my router to my system and I have both CAT8 Ethernet and fiber cables permanently in-place, so I have conveniently tried A/B comparisons between the two over a longish period.  I seem to detect a slight difference in the sound but ended up prefering the CAT8, at least with my current set-up, although YMMV.

Whether any of this changes the sound of your system, only you can decide and, unfortunately, only by listening, not by reading these forums which are great for learning what is out there but not so great for evaluating the actual sonic impact of this stuff.   In my system, none of it makes a profound difference but maybe a little so I am using a bunch of stuff, nonetheless.  IME the components you are using (i.e., streamer and DAC) have a much greater impact on the sound than cables or add-ons.  Below are the ancillaries I am using:

• LPS on both modem and router
• 45' of CAT8 into dual Bonn N8 switches with two LPSs
• muon network cable/muon filter into server
• muon USB cable into DENAFRIPS Hermes DDC
• TWL AES/EBU into DAC

Good luck.

@curiousjim 

A few years ago, Steve Plaskin (with AudioStream at the time) posted about using fiber converters and 1M of fiber as an isolation technique to reduce noise.  Based on that review, I purchased 1M of cable and two converters, and then later installed my entire 45-foot run with fiber since I figured having only one SMPS near my system was better than having both near my system.  Before long, I switched to LPSs.  I cannot remember how I came on the products that I used, and I am not the right person to ask about which specific converters would offer the best performance but these are what I used:

• converters
• fiber cable
• linear power supplies

For a while, I thought music sounded a little smooth/calmer using fiber from my router to my system, but then when I started trying other isolation products such as the Network Acoustics eno, and later the muon, the effect of using fiber seemed to be no longer necessary and I found that going back to Ethernet cable only for my 45-foot run (CAT8) may even sound slightly better, although the difference seems almost imperceptible. Your plan to start with the converters and then if you like the sound, add LPSs is a good way to start.  FWIW, the iFi power supplies I believe are highly filtered SMPSs but not LPSs.  That doesn't mean they are not quieter.  However, my understanding is that there is a big difference between the iFi SilentPower and the more expensive iFi Power elite.  Good luck.

. 

All this fiber talk made me want to hear that set-up again, so I just reconnected my 15M fiber cable from just out of my router to just before my double Bonn switches, where my gear is.  It sounds pretty good.  I doubt I could pick it out of an A/B comparison against the CAT8 cable but the impression I have is maybe a little “wetter”, or maybe a little easier to listen to.  Certainly not night and day but I will leave it in for a while.

@curiousjim 

Put the converters after your last Ethernet connection. I thought you were planning to run CAT6 to your system. The other thing uses your home electrical wiring to extend your Wi-Fi. I know nothing about that but suspect there could be a risk of adding all sorts of noise (others here know more about that than me, I am sure).  I would put the converters after the Ethernet cable that comes out of the adaptor kit, or at the end of your CAT6, if you choose that route.

BTW, unless you already own the CAT6, consider using CAT8 for the additional shielding.  The cost difference is not that much.

@jjss49 - Good point about having a LPS after the fiber conversion. I believe the OP’s intended set-up involves only a short fiber cable so both power supplies will be located near his system. If LPSs are not going to be used for both converters, then it might be helpful if the dirty side SMPS can be on a different circuit from his audio equipment, or at least not plugged into the same power conditioner (could try it both ways). BTW, the linear supply I linked in my earlier post (25W DC Linear Power Supply 5.1V 3.5A) is currently available on Prime with free returns. The OP should confirm his converters will operate on 5v since some do but others, like the converters I linked, require a 9v supply, which are not as easy to find as the more common 5v.

Finally, with all of this stuff, IME, do not expect dramatic improvements or changes in what you hear. For me, these ancillary items have resulted in subtle changes at most.

@curiousjim 

I really don’t hear any difference.

I am not surprised, and I agree with @tonywinga that you should either hear a change, or not - your perception is not likely to change over time. 

I have gone back and forth with fiber or Ethernet multiple times and as I said earlier, subtle changes at most, to the point where I doubt I could reliably discern the change, if there is any - they both sound good.

I never asked, were you trying to fix a problem or simply exploring possibilities to optimize your digital source?

@curiousjim 

So, I was thinking about getting a Denafrips Iris DDC to do the reclocking job, but haven’t made any firm decisions yet.

I bought a Denafrips Hermes a couple of months ago for the similar reason, plus to try converting USB out of my streamer to AES/EBU that I could input into my DAC.  The Hermes and Gaia offer more options for clocking, particularly if you have a Denafrips DAC, but my first step was to try it as a converter and hopefully send a lower jitter signal to my DAC. 

Regarding the sound, it seems to sound at least as good as running a USB direct from my server to my DAC.  This may say more about my server and DAC since they have quite high quality USB interface boards.  At the least, I can say putting the Hermes in the chain does no harm sonically but, again, I did not even use the clock function so there is probably considerably more potential to be had.