Having kept a close eye on the elastic hosting ecosystem and all the players that have launched offerings over the past decade, we’re happy to see some real differentiation finally developing in this space. The one we will be focused on here is that separating large publicly traded hosting providers who spend millions on marketing and their smaller privately owned engineering-focused counterparts who spend their time actively seeking out and solving complex IT problems.
What we’re seeing and experiencing is that large public hosting companies like Amazon, Rackspace, Google, etc. have taken a race-to-the bottom approach providing infrastructure as a service… sort of like how Walmart has. Though wildly successful over the past few years, and we most definitely respect and see a need for this type of service (companies such as www.iron.io and www.soasta.com for example are now doing some very powerful and awesome things with access to so much raw compute power on demand), the key attraction and business driver to this model has been that the price per compute unit will continue to fall as ever larger economies of scale are leveraged and the need to actually touch a client becomes increasingly unwanted or needed.
Conversely, the privately owned hosting companies we work with tend to take a more involved and holistic approach that is focused around delivering actual solutions rather than simply booking more revenue and shoving as many people as possible on a given platform. Not surprisingly, the smaller engineering focused companies are spawning from former AWS, Google, HP, Microsoft, CA, Rackspace, Oracle and IBM employees who grew tired of having to deploy solutions they knew would never scale, or even work, for the end user. They eventually grew tired and personally disturbed over repeatedly being told by management to smile and nod and play the game so that the revenue could be booked, sales quotas could be met, commissions could be paid, forecasts met, etc. etc..
The draw to the privately held engineering firm is that the company will give you the comprehensive personal service and support that the larger public players will not because they actually enjoy solving complex IT problems and they have their most intelligent staff working the front line so they can shoot you straight about what will and will not work early on in the engagement. This way they can send you elsewhere if in fact what you need is NOT what they can deliver, saving everyone time and money in the end. They are after the long term relationship as they understand the intrinsic value of a client successfully scaling and growing within their environment. The short term commissions earned in these firms are either not prioritized or not existent, as they are tied to the long term success of the client and the employee.
Clearly, in our opinion, this approach is not only sustainable and more personal, but quite simply the right way to do business. Equally important is that this approach is critical if network and systems engineering expertise is not something you want to have in-house but still something absolutely necessary to your success as a business.
The graph below helps us explain why:
The (%)s you see represent fractions of total spend from an average company’s engineering budget. The pyramid on the left is your traditional DevOps Model and budget. The pyramid on the right is the new un-managed model companies have started to leverage using Amazon, Softlayer, Google, and other dedicated elastic hosting providers. This graph is EXTREMELY important to understand, as the fact that you are not paying for the proactive systems & network management layer is one of the key reasons why AWS and Google can be so competitive with their pricing. It is also the same reason why unfortunately so many developer driven companies that start and stay within a non-managed elastic hosting environment stay down for days and weeks after they are unfortunately hit by an outage that takes down their servers.
While companies like Rackspace claim to provide fanatical service and support, and in truth their front line support teams are very friendly (most of the time), the reality is that the revenue focused publicly traded company offering a public/shared hosting/cloud solution would prefer to be a totally un-managed infrastructure provider (despite what you hear from their sales and marketing teams). This is because the margins are much higher when they do not need intelligent people proactively coaching clients every day on what they should or should not be doing. It is FAR easier to simply hire personable ticket takers who are trained to tow the party line and product and fit as many square pegs into the round holes available.
All this being said, if you already have the expertise to manage your infrastructure in-house, you will probably find more value with a provider that is going for the Walmart approach and who can simply offer you infrastructure on demand. Be warned however, as you must be VERY mindful of how working within a shared environment can and will affect the performance of your applications and your visibility into the raw compute power running your backend!
On the other hand, if you’re looking for an actual technology partner for your business, one that will help you scale and manage your inevitable growing pains without having to hire a full time $150K+ salaried experienced DevOps guru or team of engineers, you’ll want to work with a smaller shop that can give your backend systems the attention you require to scale and thrive. It is important to note that clearly not ALL privately owned hosting or engineering companies follow the holistic and engineering focused approach to doing business we’ve outlined above and this is precisely why we spend so much time evaluating and vetting our partners!
We hope you’ve found this post insightful and we look forward to featuring in our “Partners we Like” section a few of the privately owned needles in the hay stack we have uncovered and are referring to above.