Archive for November 2006

Peer1/BGP/load balancer updates

November 28, 2006

Hi all,

Market Post Tower has scheduled downtime tonight at 7PM for our Peer1 circuit (any IP that begins with 69.90.x.x.) This downtime is supposed to last for at least 15 minutes, and the maintenance window is 1.5 hours.

When we heard about this, we let Market Post Tower and Peer1 know that the downtime was unacceptable. We acted quickly and procured a switch with 2 GigE fiber ports. We then upgraded our BGP session with Peer1 so that we would be on a different switch port on their end, ensuring that there would be no downtime. We’re letting you know about this maintenance window as a courtesy in case something goes wrong on our end and we end up being affected by the downtime. However, we do not expect there to be any downtime tonight.

Update on getting our own IPs: After endless delays and haggling with various upstream providers, we have finally secured a fully redundant BGP link running through multiple providers (so if any single provider goes down, our network will not be affected.) The link is running through multiple routers on our end so a single router failure cannot take us down, either. We will begin switching customers to new IPs in December, and expect this switchover to take 30-60 days to complete. You will be notified directly when your IPs will be changed (please do not email us asking when you will be changed over — we’ll notify each customer individually, and we have not yet set a schedule in stone since we are still in testing mode.) We have invested upwards of $30,000 in getting this new infrastructure up and running, and we made sure to take extra time and money to get everything set up properly instead of skimping on equipment. It will be worth it!

Also, a quick update for customers on our load balancer, or those who are interested in our load balancing service: Our current load balancer had some strange problems. We discussed what we wanted to do about these issues and decided to procure a new load balancer. Like our current load balancer, our new load balancer is also Foundry brand, but has upgraded GigE ports and newer software. The new load balancer is in and we are configuring it now. We will be moving current load balancer customers to the new load balancer either this week or next week. We will be in communication with each customer individually to schedule a transition. There should be minimal downtime involved.

The times, they are a-changin’…

November 21, 2006

Welcome to Simpli Hosting’s blog! I’m going to start off this first blog post with where I intend for Simpli to be in the next 12-36 months. As you can imagine, this will be a long blog post, but if you are a Simpli customer, please take a few minutes and read it all, as it contains important information regarding your hosting account and the future here at Simpli.

I started Simpli 5 years ago, in July 2001. I started it as a side project, never imagining it to grow to be a successful business all on its own. I was patient. I didn’t take a salary until 2003, letting the revenues reinvest themselves over and over until I realized that there was something amazing going on. In fact, I had inadvertently stumbled upon a key point of success in our industry. Turns out all of us are a little isolated behind those computer screens. As humans, we crave contact with others, and sometimes no matter how much we pound on the keyboard, we need to know someone’s there, listening. Particularly when we entrust our business to another business, like many of you do with Simpli, we need to feel that there are real people behind that operation. We need to shake hands, see smiles, and really feel welcomed to continue to do business with that company.

But most web hosting companies don’t do that, and that’s really — more than anything else — an inherent “brokenness” in our industry. It’s a broken model of communication when customers are forced to submit tickets and pound more keys in order to communicate issues that should be resolved with more human contact, not less. And it’s dehumanzing to have to remember your “customer number” or your “ID” just to be able to communicate with the company behind your business.

That’s where I realized Simpli could excel. We weren’t the biggest hosting company out there. We never will be. That isn’t one of my goals. Instead, I want to make sure you get treated like a human being. And if you’re satisfied, I go home happy at night. Really.

Except…there’s a problem. In order to compete and grow, we based our pricing model on all of those other companies…the ones that outsource their support to India, or use cheap bandwidth providers, or use desktops as servers. We don’t do any of that. We say you’re getting premium bandwidth and you’re actually getting premium bandwidth. We don’t use sneaky tricks to forward your traffic out over some cheap provider. We don’t host in someone’s garage, and we use top-of-the-line SuperMicro 1U and 2U servers, not cheap desktops. Oh, and the most important thing…we support 5 full-time and 1 part-time person here in San Jose, CA (one of the most expensive metro areas in the country, where a median 3BR/2BA house is priced at $670,000.)

Here’s the point where I get brutally honest (as I am known to be!) We simply can’t afford to continue offering all of this great stuff to you at the price points we have now. Those price points were sustainable when it was just me and perhaps one other person, and I just took whatever was left at the end of the month as my salary. But those price points are no longer sustainable with 6 employees, with you–our customers–demanding more support and services. We have to do something drastic like putting all of you on cheap bandwidth providers or outsourcing our support to India in order to continue to expand our service offerings. But this company wasn’t built around “cheap.” It was built around delivering you the best service.

We’ve reached a critical point in our growth. With our first $1M revenue year around the corner in 2007, we need to decide where we go from here. We’re in an industry that changes rapidly. In the past 12 months, many new innovations have come out. But since we’re spending every cent that comes in every month, we can’t explore those opportunities. And since we’re not interested in outsourcing our support or cheaping out on providers, we’re going to raise our prices instead.

I can’t tell you exactly how much prices will go up. I can tell you that we have 24 racks full with over 500 servers total at Market Post Tower. But we’re full…we don’t have much more room to add new customers. We’re about to sign a new lease on 20 more racks in Santa Clara, CA (about 10 miles from Market Post Tower). We’ll keep our current customers at Market Post Tower and continue building out new customers and services in Santa Clara. However, the datacenter space market has drastically changed in the past 12 months, and our new lease is double the space cost of our current one. That means our prices will go up probably 50% or more for colocation, and dedicated server prices (for new servers) will increase as well.

I’d like to take the time to outline some of the solutions we plan to offer once we raise prices:

1) A better load-balancing and failover setup. We plan to hire at least one new person just to work on a solution that will keep your site up and running even if Market Post Tower gets hit with a nuclear weapon…and for under $1000 a month. However, that will take an investment of at least $100,000 to pull off. We know that if we can pull this off, it will be huge not just for us but for the hosting market in general. And we know that many of you are interested in this very solution. Automatically not only keeping your data backed up, but your website and email online even if the entire datacenter is destroyed is a great value proposition for those of you whose websites are critical to your business. We plan to bring this solution to market in 2007. But we can’t do it while charging rates that only barely keep us afloat.

2) Upgrading our billing system and integrating it with our reboot system and bandwidth graphs. We’ve heard your cries on this. We need a “customer portal” where you can access all our services and order more services. But that will take another full-time person to implement, and we can’t afford that person right now.

3) Upgrading our website. It’s circa 2002 and a complete mess right now. It’s probably the thing I dislike most about Simpli. But we can’t afford to hire a developer to put our new website up and get it running and integrated with our billing system.

4) Hiring true 24×7 staff. We hired a weekend tech, Mike, and he’s finally getting the “flow” of things and answering most support tickets on weekends. But we still don’t have overnight staff, and our current staff is overworked as is (it’s now 11:00PM on a Monday and I’m at the office…) We need to hire a team to help us answer your questions 24×7. But there’s simply no money to do that right now.

The web hosting industry is shifting. Soon the entire “dedicated server” model will become as obsolete as a mainframe is today. We need to ride that wave or risk getting shut out of a new market that develops around us. We need to be able to make strategic moves to ensure that we remain ahead of the pack (or at least with the pack) and don’t get left behind. And we need to be able to honestly keep our goal of stellar support and 24×7 customer service. My hope is that you will not only understand why we’re doing this, but also support us in our endeavor to make your web hosting experience the best it possibly can be.

I’ve built Simpli with your help, and I am asking for your help again as we transition to one of the leaders in our industry…while never forgetting that you helped us get there.

Thank you for choosing us to support your business.

-Erica