Layer42 Networks

So I’ve been a customer for about 10 years, and have gone through all of their ups and downs. Stuck through some SERIOUSLY bad service, some very slow support, and very inconvenient outages. They’ve matured in some of those areas, but as a couple of other people have posted, they are definitely lacking on contract lock-​in, inflexibility to change packages, auto-​renewal, and overall responsiveness (waiting two weeks to get an email from them about the contract).
Although their reliability has improved, it’s telling that nearly every one of their reviews has mentioned outages (power, HVAC, network, etc.). You get what you pay for, and in the event you want to move on to another provider, you’ll end up paying for months of unused services unless you luck out and make the decision to move near their 60-​day written notification window.

If you’re looking for a colo provider with great bandwidth and great support I highly recommend Layer42. My customers have hosted equipment with them for years and they’ve been top notch. Great uptime & great support. If Equinix is the Ferrari of Internet colo then Layer42 is the Porsche. Just as good but a bit more reasonable.

We have two cabinets on Layer42. Been customers for 10 months.
The onsite-​support virtually doesn’t exist. You will need to drive to the facility for every detail.
Nobody reads emergency tickets. Not single one was ever responded.
When power failed in the entire facility, nobody was answering phones for hours.
Their auto-​renewal lock-​in policy makes it hard to leave.
Layer 42 is an evil place.
update: —–
and the worst: all contracts are auto renewed, and require 60 –day notice prior to expiration. there is *only* a year option for any service. don’t phase out contracts: we had cabinets and 50mbit/​s bandwidth from august, then we bought much more bandwidth in december. since they do *only*anual contracts, cabinets were expiring in next august and bandwidth four months later. in order to leave, we’re now faced either to waste money for 4 months of bandwidth or 8 months of cabinets. no, they explicitly won’t let us buy any of these monthly.
yes, it sucks.

Layer 42 is my default choice for server colocation in the Bay Area. I’ve run colo facilities in multiple countries, and every facility has weaknesses, and is designed to a price. Layer 42 is reliable, the staff are knowledgeable and responsive, and they can provide services across the range from single servers to large cages. Layer 42 also has excellent bandwidth (transit and transport). While Layer 42 did have one HVAC/​power issue, every facility I’ve used has had at least one or two brief outages — several of the larger facilities have had multiple outages over a short period of time. Layer 42’s response to the outages (facility upgrades) was as good as could be expected.
The only area where Layer 42 is weak compared to bigger competition is «flashiness» when doing tours for non-​technical clients. Facilities like 365 Main and large Equinix facilities look much more visually impressive, and have spent much more on their equipment. However, reliability ultimately comes from well the staff at the facility operate and maintain the equipment, and even with a much smaller capital spend, Layer 42 delivers better performance and reliability than much larger sites. Layer 42 pricing easily supports colocating at multiple sites, and the reliability of two very good sites will be much better than a single «perfect» site.

Use Layer42 at your own risk. I work at a medium size start-​up with ~60 servers in a cage at their facility. Initially, things went pretty well. The last month has been a nightmare.
Not more than 3 weeks ago the power went out to their cooler, slowly raising the temperature of the entire colo, on a Saturday no less. Our internal email thread looks like the SOS stream from the Titanic. Nothing to worry about at first, just a few servers down. As the heat rises in the colo, more servers fail until we decide to power everything down ourselves to prevent complete catastrophic failure. Did I mention we run a fairly large ad-​network and for every second we are lights out, we (and our partners) lose money? Fun! The best part is when our ops team shows up on prem and the colo manager says «Don’t worry, we lost power for everyone, not just you guys». Huh? Well, shit… thank god it wasn’t just us.
Last week, Layer42 has a «brief» power outage. Just a couple of minutes. Let me tell you what a sudden and complete power outage feels like. Imagine you’re walking down the street and a random stranger stops in front of you, pauses briefly, then gives you a huge kick in the junk. Sure, it’s only a brief moment of outage as your roll on the sidewalk wondering what the hell happened, but the lingering effects last for days as you get «everything working again».
How can you prevent this from happening to you? First make sure your SLA has specific coverage (with penalties) for any power outage. Time based outages don’t properly measure lost revenue or time as you piece things back together again. Second, absolutely have a multi-​colo configuration with instant fail-​over/​BCP.
Going with 2 stars instead of 1 only because they place hasn’t caught fire or otherwise imploded taking our equipment down with it. That will probably happen before the end of the year though, so check back later.

Layer42 has my vote for best colo ISP in the Bay Area, I used to be the CTO of a competitor, now I’m a customer; I think that says a lot.
I have a decent sized cage out there and I’ll be expanding soon.
Layer42 has the flexibility I need and the ability to deliver with no hassle.
Seriously, Layer42 is going to surpass the competition soon.
— Update —
I saw Jason M’s review and I felt like I should offer my perspective.
First and foremost, I have yet to see a significant HVAC outage at Layer42’s Datacenter…
Second, I experienced the same power outage at Layer42, and in terms of things, I’m still very very happy. Datacenters do occasionally lose power, even if they’ve spent millions and millions of dollars to prevent that very thing from happening. Almost every datacenter I’ve ever been in has lost power at least once or twice. It happens, you have to live with it and plan for it. If you plan for your datacenter to be up 100% of the time, you should reexamine your plans. You also shouldn’t load up your PDU’s above rated capacity, because that will take you down faster than any datacenter related outage.
And in a non-​JasonM related note, because of my leveraging of Layer42’s facilities I’ve grown my business very very quickly and in one year I’ve surpassed where my old company was in their third year in the same line of business. Choosing Layer42 was obviously a winning decision for me :)

Horrible service, poor response, no dedicated on site tech and auto-​renewal of your contract term make for a horrible colocation company.