The perception part comes from Softlayer’s advanced admin toolset and the generally slightly lower latency to the Dallas data center. This is minor, both hosts are an improvement over our self-hosted T1 for most players.

“Good enough” may not have been the best choice of words for readers of the blog who don’t know me. Anyone who does know me (most of the readers are players of the game) already know that “good enough” is actually pretty damn good, because I’m picky like that

It’s not just about being elite it’s about providing what you guys want, and we’re always looking to improve our services to maximize your satisfaction.

Have been meaning to post a followup article to this, but the brief version is that we went with our test instance and website/mail server/etc at Liquidweb and the live server at Softlayer. So far (and it is early days in terms of judging long term stability) I have been very pleased with both.

Softlayer is more expensive (by around $100 per month on comparable setups) but their user tools are fantastic. The VPN in particular is excellent - I can connect to the softlayer VPN and get a direct console to the server on a 10 subnet address. The dashboards are more mature and they have full APMI monitoring. I can even mount a rescue disk remotely and boot to that then boot back into the normal server when I’m done. This may all be possible at LiquidWeb, but if so it is not very visible.

In terms of latency, around 200 users helped me by before we chose by sending in pings to both servers from all over the world and, on average, ping times to SL were lower, but both were excellent.

My (mostly anecdotal) perception at the moment is Liquidweb for “good enough” hosting / stability on an important service and Softlayer for the higher end critical stuff. We’d have probably been fine taking both servers with either, but I like the redundancy. If one of our ports are ever down the other should always be up. They both blow away LayeredTech btw, but that’s a whole different story (rant) for another day…

Thanks,
Dave

1. Does the site have multiple diverse routing for data ingress? In other words, I wanted to ensure that Quest (for instance) entered the building from the north, and AT&T from the south so that if one carrier’s SONET ring got backhoe’d the other seamlessly ensured uninterruped access.

2. I always visited the site and spent a day shadowing their tech staff, ensuring that I established a relationship with both the help desk manager, and NOC (Network Operations Center) people. If they weren’t organized enough to assign a real live breathing tech to me so that I could meet them, guage their competence, and buy them lunch… they weren’t organized enough.

3. Proximity is important, but not The Deciding Factor. If the data center is close enough for you to drive to it over the weekend and tinker with the hardware and/or software configuration (and their policy allows you to be “hands on” while retaining responsibility for the system), you’re going to be alot happier than if you have to fly to Denmark to try out a new NIC/driver config.

4. Ask for the last six months QOS (Quality of Service) report. If they aren’t at least four nines (99.99%) reliable, move on. Similarly, you want to talk to a minimum of five of their other customers and get their take on how good (or evil) the hoster is.

5. Environment: Fire suppression, how many emergency generators and how many days of fuel onsite, how often the generators and bus transfer mechanisms are tested (if they don’t shift the load at least monthly, move on), rated hours for the facility UPS, Air conditioning load, security, 24 hour manning, yada.

6. Don’t be afraid to negotiate. Incredibly, I found that every single data center/hosting facility had published rates and that people actually PAID the published rates. My strategy was to show up, meet everyone, make some friends, trade emails for a week, build anticipation with the salesman I was working with… then say I’d love to do business with them, but I just got a quote (and I made sure that I really did have a quote in hand) that was 20′ish% lower than theirs… what can you do for me? Now, AARD is considerably smaller than the dozens of servers I had, and your leverage may not be as convincing; but nothing ventured, nothing gained.

Good luck! Fun times!! –Flat/Odd

#5: Mirara, shouldn’t need to open them any earlier than was already planned. Costs are a little higher hosting two larges servers, but comparable. SoftLayer is significantly more expensive than Liquidweb for the same spec, but does seem to have the edge based on ping times and the peers into their Dallas data center.

You going to be running donations for this?