We’ll make this right for you.

— Corinne Moore, our Peer1 Account Manager

Since I built Status Shuffle 4.5 years ago, I’ve used five different hosting providers to run the core application and three other providers for auxiliary servers.  My experiences ranged from truly horrible (as with Peak Web Hosting where we terminated service after 1 week) all the way to excellent.  On the Excellent corner we have Peer1 and SoftLayer.  Let me tell you why.

The Five Pillars of an Excellent Hosting Experience

When I started out, I had no money so cost was the #1 driving factor.  Low cost doesn’t necessarily mean lower value, just as higher cost doesn’t automatically mean better value.  In fact, I can say that over the year I’ve gotten damn good value for my money with various budget providers (and vice verse).  Nowadays I look for the best blend of the following:

  • Cost: Still important.  You don’t want to overpay, you want to pay just enough to maximize the next four.
  • Physical location: The physical location of the server matter for a variety of reasons:
    • Distance from your end users: The closer the server to your end users (as judged by millisecond latency) the better their experience is.  If you’re building a game server for people in Australia, then it better not be located in Florida.  Beyond the physical distance, your provider’s peering agreement can also mean a short route to users nearby and a long one.  I’ve seen traceroutes to smaller providers go from Toronto to Chicago and back just to hit servers that were 5 miles away from me.  Always test!
    • Distance from an API provider: If the purpose of your server is to work with an API, then the latency to the API provider really matters.  For example, if your server needs to make a lot of requests to Facebook, then you better be near Facebook’s servers.  If your server is part of Google real-time-bidding exchange, then you your best bet is the next rack over.
    • Distance from you: The closer the server is to you (latency wise), the better your personal experience is when you do “stuff” on your server.  I always like my stuff to be done faster.  But more importantly, if the server is close to you physically, you can do some of your own maintenance as opposed to relying on 3rd parties.  Those 3rd parties may be capable, but they don’t know your infrastructure nearly as well as you do.  Personally, knowing that I don’t have to get on a plane to access my mission critical servers helps me to sleep better at night.
  • Uptime: What percentage of the time are your servers working and reachable?  With one provider, it seemed like there was always something happening.  I knew things were really bad when I noticed that I knew their tech support number by heart.
  • Time-to-fix: When something goes wrong (eventually something always goes wrong), what’s their ability to diagnose and recover from the problem quickly?  And I don’t just mean network connectivity: If your servers are managed, the provider’s ability to upgrade/fix parts quickly and efficiently are a big part of the story.  I had one provider where uptime was great, but when the shit hit the fan, it really hit the fan.  It got to the point that if I needed upgrades to the server I was scared to death of prolonged downtime (and rightfully so).  That same provider at some point had a big power outage and it took them 36 hours to get my server online after power came back.  Adiós1.
  • Quality customer support: You server is down.  Customers are calling you every two seconds.  Your sales partners are chewing your ear off.  Everyone wants answers.  You want answers, damn it!  You call the support line, all nervous and panicked.  This is money time, and customer support can blow it in so many ways.  For example, they can:
    • Ask you to “file a ticket” and refuse to let you talk to the tech on the floor even though you have pertinent information knowing full well that ticket won’t be seen until everything is long over
    • Refuse to provide, or have no access to, any useful information about the state of things
    • Blame you first, investigate later
    • Call you in an attempt to save your business only after you call to terminate your hosting due to any of the above.  (Note to all service providers: Customer retention efforts are best when preventative, not post-mortem).

A Shout-out

I’ve lived through every negative example listed above, but never ever with the following two companies.  Here is that positive feedback I promised in the last post.

Peer1

I’ve used Peer1 for a long time.  I can only recall two times where the network became unavailable and in both cases it was resolved in under 10 minutes.  (Uptime: Check, Time-to-fix: Check).  A couple of years ago I wanted to move servers from LA to Toronto so they would be physically closer to our base of operation.  Luckily Peer1 is in both cities (Physical location: Check).  Pricing was reasonable (Check.  Though what’s up with the 1 year minimum commitment?  Please stop with BS).  But what really won me over was the customer service.

Peer1 shipped my servers from LA to Toronto.  When they arrived, I drove down to rack them.  I opened the boxes and to my horror I was staring at two dented servers.  Whoever packed them in LA did a horrible job, and although four servers were fine, two looked like they were hit on the side with a hammer.  The UPS looked like I dug it up from a dumpster.  I don’t need to tell you that servers don’t like getting smashed.

I literally froze, thinking about the time and money cost of replacing those servers.  Then I called my customer rep.  Now, there are many things she could have said.  99% of those things would have caused Peer1 to lose me as a customer for life.  But what did she say? “We’ll make this right for you.”  That’s it.  Pure and simple.  They fucked up, and they will fix it.  If anybody at Peer1’s reading this, you need to give Corinne a raise.  I’m still paying you thousands of dollars a month, you can use some of that.

Customer service: Check!

In the end the servers were miraculously fine, and Peer1 paid for the damaged UPS.

SoftLayer

I was referred to SoftLayer by a friend.  Like anything else in life, it’s easier to choose when you get a recommendation from someone you trust.  (If you’re evaluating a provider and don’t personally know anybody who’s using them, try searching Web Hosting Talk forums, but be prepared for negativity bias and obvious astroturfing).

As with Peer1, network uptime is pretty good, and recovery is very fast when there are problems.  Reaching technical support is easy and they are very friendly.

One nice surprise came during a hardware upgrade I scheduled for 4am.  I was trying to call tech support, but accidentally called the sales line.  And someone picked up.  At 4am on a weekend.  “So what?” you say.  I’ll tell you what: With every other managed hosting provider I worked with, when you’re having problems at 4am and need some sort of hardware change (e.g. more RAM, more hard disks) they may refuse to do the work until Sales come in on Monday and price it out.  I can still remember myself screaming “I’ll work out the fucking price with Sales on Monday just get the server working now!” at a previous provider and believe me it didn’t sound that nice when I was actually screaming it.  So obviously, I like the fact that SoftLayer has their act together.

To top it all off, the prices are great (through you do have to call in and negotiate), no minimum commitments (yey!) and they get you up and running super fast.  That 4am hardware upgrade didn’t cost me a cent, even though it involved a tech off-hours.  All servers have remove KVM built in which can be a life saver, and you don’t pay for traffic flowing to your server (only traffic flowing out).

Highly recommended.

 

——

1 In reality moving providers is a very painful task, involving a search for a new provider, migration planning and testing, setting up new servers and the possibility of downtime when moving.  So “adiós” not only means you’re leaving, it also means you’re screwed.