There is one thing in common in Silicon Valley among all the companies and startups I've visited: young, energetic men and women trying to take over the world.
But with such youth, CEOs of companies are coming across a problem: lack of experience.
Personally, I've been in the "Internet industry" since before the world wide web became popular. When I started using this thing called the "Internet", I was using a Gopher client, running on a VAX system, to browse from network to network. What is Gopher you ask? It's the text-only predecessor to the web browser. Yeah, that long ago. But ever since then though, I've been tinkering with networks, routers, servers, operating systems, and all other kinds of things related to the Internet, basically 7 days a week.
Here at ServerPoint, there are several of us "old schools", as the young ones like to call us. When presented with a problem though, it is very easy for us to think of a solution: any problem that exists today, it is a problem that has existed in the past in some other way. It is the reason why I can grab a modern video game and easily beat it, even though I've never played it before. Why? Because modern games are just remakes of ideas of games from decades ago, just in a more modern format. This surprises my niece and nephew every single time.
There is, unfortunately, a side effect to this lack of expertise in Silicon Valley.
A CEO recently asked me during a lunch about how to reduce their AWS bill (at over $650,000 per month). His lead engineer was there and gave me some details of how they use their AWS "instances". When I realized what they were doing, I said "we have a customer with similar usage, and their bill with us is $90,000 per month". Perplexed, I explained how that can be possible. In the end, he wondered "why none of our guys have ever suggested this before?".
Our client is running a hybrid approach: they use Google Compute for some tasks, but the intensive stuff, a large image processing and storage operation, is running on a set of 175 dedicated machines at one of our data centers. The prospective client I was talking to though was doing it all on AWS, which is many times more money for the same performance.
The answer to "why none of our guys have ever suggested this before?" is simple: very few "old schools" in the industry. Silicon Valley is hiring young talent left and right and to most of them, the only thing they have grown up to know is AWS; they haven't been in the industry long enough to know there are other alternatives.
Now, don't get me wrong. The scalability of the big companies come in handy for many companies. If you are the size of an Uber, Netflix, or Apple, it is hard to grow without access to AWS, Google Compute or Azure. But even these big Ubers and Netflixes use a hybrid model: they use the big cloud companies to handle surge, but they all also have their own dedicated infrastructure somewhere else:
Apple uses the public cloud, but also uses their own data centers as well as private infrastructures hosted at third party data centers, such as Equinix. Netflix uses the public cloud, but has built their own content distribution infrastructure hosted at hundreds of Internet Service Providers and third party data centers. Uber uses the public cloud, but also operate their own servers collocated in third party data centers, such as Equinix.
We have built several hybrid models for clients, with our private cloud solution hosted in Santa Clara, in third party data centers where Amazon AWS also hosts many of their own servers for their US-West region, and at several data centers in the Ashburn, Virginia's "data center valley", which also host many of AWS' US-East region infrastructure (they use over 40 third party data centers there to host their infrastructure).
But this hybrid infrastructure is simple: hundreds of servers, running dual Xeon E5 processors, or the newer Xeon Scalable, with our ColossusCloud software running on them to perform virtualization, and direct links to your Amazon AWS private network.
These clients, using mostly containers managed by Kubernets, easily shift tasks back and forth between AWS and our private cloud servers.
Want to cut your AWS bill in half or less? Contact us! We can build one of our private cloud pods at any data center you desire and plug it to your AWS private network, starting at $49,000 per month. Sounds like a lot? That same level of power at AWS is over $300,000/month!
Ask for me, Peter Cruz, when speaking to any of our customer service or sales representatives. I'll cut down your public cloud bill!