What Does it Cost (Part 3)

January 27, 2011

Determining the value of "On-Demand"
It's 2011, and as we bookend the tail part of 2010 and the beginning of 2011 in effort to close strong and get a good jump on the year, it can be easy to lose sight of the big picture. As it turns out, this distracted me, so it's been a while since I cranked out the previous installment of this "What Does it Cost" series.

As a quick refresher, the idea behind this series came from listening to keynote speakers in conferences this past year who harped on the necessity of getting more value for the same (or even less) budget. In What Does it Cost (Part 1), we discussed how opportunity costs are the most overlooked and important part of planning an infrastructure. In What Does it Cost (Part 2), our focus was on how your people relate to your infrastructure.

The goal of this series is to fairly assign a value to what a company like SoftLayer provides relative to the costs of doing it in-house or by using colocation.

Let's start by making sure we know what 'On-Demand' actually means: 'On-Demand' means that you get what you want when you want it and for the time it is needed. No more and no less. It pretty much takes out all opportunity costs. On–Demand is good, it is necessary, and in the future it will be the difference between successful businesses and ones that are destined to fail.

Everyone has heard the term that "time is money." Receiving server and CCI infrastructure without delay when you want them they should be valuable, right? But just how valuable is that delivery? Have you ever wished you could go back in time and change something about your past? I think we all have. If I went back and took a "risk" in the stock market, knowing what I know now, I'd be writing this blog on a platinum-plated computer. But betting on a game you've already watched isn't really risk.

Infrastructure investments are risky. In some cases, their reward will justify their risk, while in other cases, taking a risk-averse On-Demand approach provides the best outcome all around.

It's tough to assign a once-and-for-all value to On-Demand options because it is not only different for every business, but it is also different for the specific scenario that you are provided. An idea is abstract and meaningless unless we can apply it in some kind of practical application, so perhaps the best way I can illustrate the decision between going an On–Demand route is by looking at it through the lens of two scenarios:

  1. Your project your company will grow by a factor of 10 over the next years.
  2. Your company is doing a hardware refresh for an advancement you want to take advantage of.

Growth
If you have to plan these things out ahead of time, it means making commitments and spending upfront capital and time in order to get what you need in order to grow your infrastructure to meet your projections. There is a lot of uncertainty and risk. What happens if the market takes a sudden downturn and your infrastructure needs to also adapt quickly? Everything works out fine as long as the future goes according to plan, but what if the projections we got were wrong?

Even if we grow 50% (which could still be a huge feat), the fact that we planned for 1000% growth could leave us financially crippled. What if the opposite of this scenario happens and your projections underestimate your future needs? This would seem to be less risky, but in reality, not having the tools necessary to provide your services or support your clients can be even more devastating. Never underestimate the cost of not being able to deliver and keep up with demand.

Hardware Refresh
For a hardware refresh, chances are that there is new software available and new ways of doing things to maximize the potential of recent advances in hardware. Generally, I see that companies that lean away from upgrading to "new" and simply keep the status quo. Why? The risk is too high, and time invested by personnel is far too costly. It may not be appealing to take the risk on finding what will make a new solution work when the cost of that investigation is high and the results are uncertain at best.

The problem is that if you stay unaware of changes in technology you'll soon find yourself getting further and further behind. The thought of maintaining the status quo can be as dangerous as quicksand.

Comparing In-House v. Outsourced
OK, now that the scenarios are set, let's look at what would happen when we play out and in-house infrastructure vs. an outsourced On-Demand solution with SoftLayer.

In a data center environment, the decisions you make have a long term impact. Once you make your decision and spend your money to host in-house, you're fully invested in that decision. The only thing that could break from your plan is a catastrophe. The sensible thing to do is to take time to better educate yourself so you can make better decisions. That means businesses will regularly take months (I've even heard of instances taking more than a year) to devise a strategy and months more to execute the implementation of that strategy. As a result, companies try to plan in multi-year cycles (5 years seems fairly typical).

Think about this: What has happened to your business and in your own personal life in the past 3 months? Do you feel confident in predicting exactly what will happen in the next 5 years?

While SoftLayer might not be in a better position to predict what will happen in the next 5 years, we operate an 'adjustable' infrastructure, so we'll be ready for whatever may come. Instead of an upfront capital expenditure, you can pay monthly to get the newest innovations in server hardware, and you can upgrade at any time without penalty. Instead of signing the long term contracts inherent in running your own data center environment (space, power, bandwidth, software, etc.), you can have hour-by-hour terms, month-to-month at the longest.

Running your own facilities means waiting weeks/months for your hardware to arrive so that you can have it racked and put into production. SoftLayer can build you customized dedicated server configurations that can be provisioned in under four hours. Cloud Compute Instances (CCIs) can be added in minutes, and by using templates you can save even more precious work time. You can even go as far as to automate this by utilizing our API-driven customer portal.

Even if the future doesn't go according to plan when you're using an outsourced On-Demand provider, you will have succeeded in eliminating much of your long-term risk. You can make the necessary adjustments to keep your business in the best position, regardless of what happens.

To give this example some teeth I'll tell you about a customer that I recently assisted: Customer X was looking at adding four fairly stacked servers and a SAN Solution to their infrastructure. To manage that infrastructure in-house, they determined that they would need to add two employees. All in all, this was going to cost them $140K in upfront capital for the hardware, and $120K per year on personnel (if they were lucky). This was all before they could see if their current in-house data center environment could support the additional infrastructure.

As it turns out, the data center environment couldn't sustain the power, and they would be forced to re-up on three year-term contracts for more space, power, and bandwidth to move their existing infrastructure to a larger portion of their data center. This project's costs were getting out of control, but they needed to make a change to deal with business growth. The problem with executing this plan is that at the end of the day, the business growth might not be able to justify the cost of expansion.

The worst part of this is that they were "pot committed" (for any readers that play poker) because of the big-money deal on software licensing they had already executed.

SoftLayer helped them by offering different 'on-demand' ways they could get the job done. As it turned out, they were ordering enough hardware to plan 18 months out and they expected further growth on a longer time line. They were not planning around what their hardware needs were today ... which was really about a third of what they were planning to purchase in the short term. We were able to set up dedicated servers, integrate Cloud Compute Instances for short term spikes in CPU needs and work in a storage solution that could be grown as their needs increased. To top this off, we also developed an High Availability (HA) strategy that put pieces in place where they could easily shift their entire operation into a different data center in a different city, should it ever be necessary. This was an added value that they knew they couldn't come close to executing themselves.

The best part of this example is that even after about a year of service and maintaining a consistent growth pattern, they still have not spent what they would have just in paying the additional two employees they would have had to hire. SoftLayer gave them the means to save $140,000 up front and thousands per month ever since. That customer is planning on moving the rest of their infrastructure into our facilities when their current contracts run out ... They've told me every time we've spoken that they want to make this move immediately, but they are still paying for decisions they made years ago.

At least for them there is light at the end of the tunnel, and they will be truly taking control of their infrastructure.

-Doug

Comments

January 27th, 2011 at 11:17am

I would like add our experiences with SL to this piece which further supports their blog.

We have been with SL for 2 years now with 2 boxes, and the few times we have had Centos challenges, every person I spoke with via the trouble ticket system, or via phone, were extremely knowledgeable, courteous, and fixed our issues!

We shopped quite a few hosting companies a couple of years ago trying to find a new home, and even though SL was more expensive than some of the "competitive" offerings, the service is what puts them them over-the-top, and makes that monthly invoice worth it.

Ever since moving to SL, we sleep easy at night. Inside of 2 years, we have yet to experience downtime due to SL's network.

Go SL!

Regards,
Eric Caldwell
CEO
Web-JIVE.com

February 11th, 2011 at 2:30am

Hi Doug,

I enjoyed reading your post and I wanted to add that in addition to the cost savings which a customer can obtain through the virtualised datacenter services which SL provides, the other key reason which I believes makes SL different from other datacenter providers, is your committed and proficient team of account managers, support staff and management team which has seldom, if ever let us down.

Inspite of being many thousand of miles away from your datacenters, I have never felt any perceptible delay in receiving efficient support from your team which is a feat not many providers can consistently provide on a long term basis.

Keep up the good work guys!

Cheers,
Hriday Biyani

Leave a Reply

Filtered HTML

  • Web page addresses and e-mail addresses turn into links automatically.
  • You can enable syntax highlighting of source code with the following tags: <pre>, <blockcode>, <bash>, <c>, <cpp>, <drupal5>, <drupal6>, <java>, <javascript>, <php>, <python>, <ruby>. The supported tag styles are: <foo>, [foo].
  • Allowed HTML tags: <a> <em> <strong> <cite> <blockquote> <code> <ul> <ol> <li> <dl> <dt> <dd>
  • Lines and paragraphs break automatically.

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.
By submitting this form, you accept the Mollom privacy policy.

Comments

January 27th, 2011 at 11:17am

I would like add our experiences with SL to this piece which further supports their blog.

We have been with SL for 2 years now with 2 boxes, and the few times we have had Centos challenges, every person I spoke with via the trouble ticket system, or via phone, were extremely knowledgeable, courteous, and fixed our issues!

We shopped quite a few hosting companies a couple of years ago trying to find a new home, and even though SL was more expensive than some of the "competitive" offerings, the service is what puts them them over-the-top, and makes that monthly invoice worth it.

Ever since moving to SL, we sleep easy at night. Inside of 2 years, we have yet to experience downtime due to SL's network.

Go SL!

Regards,
Eric Caldwell
CEO
Web-JIVE.com

February 11th, 2011 at 2:30am

Hi Doug,

I enjoyed reading your post and I wanted to add that in addition to the cost savings which a customer can obtain through the virtualised datacenter services which SL provides, the other key reason which I believes makes SL different from other datacenter providers, is your committed and proficient team of account managers, support staff and management team which has seldom, if ever let us down.

Inspite of being many thousand of miles away from your datacenters, I have never felt any perceptible delay in receiving efficient support from your team which is a feat not many providers can consistently provide on a long term basis.

Keep up the good work guys!

Cheers,
Hriday Biyani

Leave a Reply

Filtered HTML

  • Web page addresses and e-mail addresses turn into links automatically.
  • You can enable syntax highlighting of source code with the following tags: <pre>, <blockcode>, <bash>, <c>, <cpp>, <drupal5>, <drupal6>, <java>, <javascript>, <php>, <python>, <ruby>. The supported tag styles are: <foo>, [foo].
  • Allowed HTML tags: <a> <em> <strong> <cite> <blockquote> <code> <ul> <ol> <li> <dl> <dt> <dd>
  • Lines and paragraphs break automatically.

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.
By submitting this form, you accept the Mollom privacy policy.