Opinions? 10GB SFP+ to 10GB RJ45 Converter GBIC

Notice: Page may contain affiliate links for which we may earn a small commission through services like Amazon Affiliates or Skimlinks.

John Titor

Member
Jul 20, 2016
123
17
18
United States
Hey all, I'd like your opinion on this GBIC and to know what you guys think as well.

The GBIC is HP branded and came out this year. It goes into a 10GB SFP+ port on either a switch or an SFP NIC and converts it into an RJ45 ethernet port. Now I know that 1GB converters are extremely common, but this one is 10GB to 10GB, which in my opinion is a game changer. So far I haven't seen these from any other manufacturers and when I asked some old colleges about this, they hadn't heard of one of these yet either.

Essentially, this is extremely helpful if you have an SFP (10GB+ per port) ToR switch in your rack, but you happen to have some 10GB RJ45 NICs in your servers that you need to utilize (or you need to connect your ToR switch to a 10GB RJ45/Ethernet switch somewhere else). Before this, you'd either need to use a 1GB converter (losing 90% of your bandwidth), or you'd have to fork over the money and rack space to put in another ToR switch with RJ45 ports instead of SFP (or for smaller projects, just change or add a NIC to your servers).

Now here's where it gets all fuzzy and confusing to me... HP charges roughly $1000-1300 per GBIC. Yes, a single GBIC that costs over a grand. Before I continue, check out this link where I posted an album of pictures of it. When I saw something this small and simple associated with such a huge price I just had to ask others about this...

Pictures!! Click here -----> 10GB SPF+ to RJ45 GBICs

So obviously this device is extremely convenient to have, but I can't see how it would make sense for 95% people to buy one. Clearly you would just buy a different NIC for your server, or if you need lots of both types of port, you'd just buy a second switch for around that same cost, right? So my opinion is that these are only useful for seriously large enterprise-grade data centers where each rack's U-space is extremely valuable and they "need" to have this sort of flexibility (no matter the cost).

My Questions
- Do you guys agree that only large companies would even consider buying these?

- Have you ever heard of these before AND/OR know of any other manufacturers that are selling these?

- Do you think these are so expensive due to their rarity? (I would think once the market is flooded with them, they'd be much cheaper...)

- Lastly, do you think these will ever have a place in the home-lab environment?

I know this is is probably really boring to many people, but I had been using several of these at my previous job for the sake of convenience (no, I didn't suggest buying them). When I learned how rare/expensive they were, I was honestly fascinated since I thought they were common-place already. I've got a few of these at home, but now I'm considering contacting an HP re-seller to sell or trade these for a switch.

I'd actually really love to see a review of these on STH if people were interested. (I'd totally be willing to send the hardware to Patrick if he were interested)
 

xhypno402

New Member
Dec 7, 2016
10
2
3
41
There are a 2 big downfalls of these.

1. Limited range due to reduced power of the sfp+ port.
2. Price

I don't see big enterprises buying these as 99.99% of them go fiber. I would see this as more of a SMB/MMB thing where they have servers from the newer intel c612 chipset with it that are including 2 x 10gbase-t ports on board. I would see in this case if they are not planning on running more 10gbase-t and do not want or have the pci-e space to add an add on nic and already have an sfp+ switch.
 

OBasel

Active Member
Dec 28, 2010
494
62
28
i'm still excited about qsfp to sfp today. 10g copper and sfp+ are very different markets. copper = old data centers and those that have been running fine on 1gbe. anything newer and you're going to see sfp+.

my first stab at this problem is that sfp+ spec prob has a power budget and doing the signal processing for 10g is too much for that tight of power envelope.

we'll never see a flooded market of them. if people were making tons of them, fs.com and other companies that just are overrunning chinese factories to make cheap optics would just do the same for these converters. without many on the market
 

John Titor

Member
Jul 20, 2016
123
17
18
United States
@xhypno402 & @OBasel Thanks for the comments. Definitely interesting to hear your thoughts.

With all of the less expensive alternatives it still baffles me that they would even produce these with this cost. Kind of seems like wasted effort on HP/HPE's part to go through producing these from my point of view if the markets are so narrow and there are so many alternatives. It still seems too expensive to fit the exact scenario that hypno is describing, along with how narrow the chances might be to fit that mold. (I say that because even those companies could just buy an 8 or 16-port Netgear 10GB ethernet switch for the cost of one of these, right?)

@Patrick have you ever seen one of these? Do you see any reason this might make sense, especially given the cost? When I wasn't aware of the cost people put on these, I actually imagined these to be an amazing addition to the home-lab type of environment.
 

lucidrenegade

Member
Dec 12, 2011
75
20
8
If you're an enterprise with all SFP+ switches and for some reason you need a couple of 10Gbase-T ports, these are a whole lot cheaper than dealing with switch replacements or adds. Most large enterprise network engineers aren't going to want to throw a Netgear switch into their Cisco/Arista/Juniper/HPe/whatever environments.
 

John Titor

Member
Jul 20, 2016
123
17
18
United States
Looks like they only have a few left. That and it doesn't show any HP/ProCurve compatibility for whatever reason.

I think my best bet is probably just to throw these on eBay and then buy a switch with the earnings. It seems like a bit of a stretch, but there just might be someone out there that fits the minuscule criteria of people who would want these.
 

aero

Active Member
Apr 27, 2016
346
86
28
54
I really see no use for this kind of transceiver at all, unless one is in an environment of hodge-podge crap with no budget to buy proper equipment. And given the price of them, one could buy said proper equipment!
 

cheezehead

Active Member
Sep 23, 2012
723
175
43
Midwest, US
If your looking at those in the datacenter, I would honestly just look at using DAC cables. However, if your planning on running a more than a few runs.....save yourself the headache and just go fiber. 3rd party short-range 10gb SFP+ modules are under $20/unit these days and the glass is fairly cheap. Remember, you don't need branded SFP+ modules...."compatible" yes but they can be third-party.

I could see using these or more likely a 3rd party compatible version for workstations if you already have the CAT6 (or better) structured cabling in place (assuming your within distance limitations.

At home, I could see using something like these for my desktop but for my homelab just using glass (pref) or DAC cables because they are in general cheaper.
 

John Titor

Member
Jul 20, 2016
123
17
18
United States
@cheezehead I fully agree that for almost all the use-cases there's a much cheaper alternative (like just running fiber, using DACs, etc.), but I'm still trying to figure out why HP/HPE would bother making these, and more importantly, why make them so expensive.

I guess in the end it really is probably just for those extremely tight fits in enterprise racks where they absolutely cannot substitute a NIC in there servers or U-space at the tops of their racks.
 

cheezehead

Active Member
Sep 23, 2012
723
175
43
Midwest, US
SFP modules have a really good markup, each vender does there custom "coding" on the modules so they only work in their gear. 99% money grab, 1% actual compatibility issues. We keep around a handful of HP-branded SFP+ modules just in case support gives us grief but 99% of our modules are sourced direct from China. I just can't fathom why I would want to spend $1,500/module when 3rd party modules can be picked up for under $50/module. Is there a higher failure rate with them?...yes there is, once we had a bad batch and every module in the batch was basically DOA...after they've been up and running though we haven't seen any of them fail in years.