5 Cloud Predictions: Where’s it headed? What’s the future?
Interestingly enough, it seems I’ve had the same conversation 4 or 5 times in the last couple months. More or less it all starts the same way: “What do you think of the acquisition of X?” or “Are you worried about Y entering the cloud/hosting market?”. I believe if we look far enough back at how the Internet, Web Hosting and IT in general has evolved we’ll see that it’s pretty easy to understand.
As a business owner of a technology based company, I have the difficult job of having to be aware of where business is headed and as the same time where technology is headed and align the two such that we’re there early enough to make a reasonable profit! So here’s what I believe is happening in our Industry and where the cloud is headed.
#1 There will be no more Web Hosting servers!
What do I mean? Let’s look at the history of what I’ll call “the commercialization of the Internet” in a 30,000 foot view. The Internet is basically:
- The Bandwidth/Pipe to the customer (early on these were the dial-up lines and the T1’s)
- The Datacenter (early on this was the closet at the ISP and later gave way to carrier hotels)
- The Physical Server
- The Server OS Instance
- The Website
Earliest web hosting companies were local ISPs. These ISPs owned all of the elements of the web. The connectivity (the pipe) to the customer (you know, that 28.8K dialup connection and the T1 upstream to the Internet), the datacenter (in the form of their server closet), the servers (in the form of some whitebox PC they built), the server OS (generally some form of Linux back then), the website/workload and they hosted each website on a dedicated server.
Over time, larger ISPs came in and no longer did the ISP own the pipe and the server. The pipe went to the ISP and the server to the web hosting provider. One could say this was the birth of web hosting and it was at this time that the web hosting provider no longer needed to own the pipe to the customer.
As technology evolved and businesses started to rely on their website and email as mission critical we found the early web hosting company that had a closet or at most a server room, giving way to larger datacenters. Either hosting providers were building large regional datacenters or they were leasing colocated space in carrier hotels like the one we use (the NAP of the Americas) where the economies of scale were realized to make this level of infrastructure more affordable. It was at this point that hosting providers no longer owned the datacenter.
As web hosting continued to evolve companies like Rackspace, The Planet / SoftLayer, LayeredTech and Peer 1 started to arrive and pure play shared web hosting companies no longer found it necessary to own the physical server either! A great example of this is HostGator and the dozens of companies that have followed them. HostGator is one of the largest shared hosting companies in the world and to the best of my knowledge they still rely on dedicated servers from providers like SoftLayer. So here’s a multi-million dollar hosting company that doesn’t own the pipe, the datacenter or the server. They do however maintain/own the server OS and the website/workload and thus the customer.
So over time, hosting providers no longer owned the pipe to the customer, hosting providers no longer owned the datacenter and they no longer owned the servers. Then it stands to reason that as this evolution progresses, will we even need to own the actual server OS instance?
I believe this is where the cloud comes to play. I believe we will own the customer, we will own the website / workload and we *might* own the actual OS instance that the customers run in but this instance will, in many cases, run on top of someone else’s cloud infrastructure like Microsoft Azure.
#2 There will only be 4/5 worldwide clouds!
As we continue to evolve we’ll find there will be a number of clouds but only a handful of global clouds. I believe Microsoft Azure will most definitely be one of these as it’s critical to Microsoft’s business to be the #1 cloud option. I believe Google will have some sort of cloud then the question is who will be the others? Well, I don’t know but I’d say most likely Amazon based on the innovation and progress we see from them and *possibly* Rackspace. These will be worldwide cloud providers and on top of these clouds the classic web hosting workloads will run. We’ll find the next generation of HostGator companies running here. But it won’t be just these providers. You see, it’s pretty obvious that because of International laws, cultures and what not we’ll have to have regional providers as well.
We’ll find that you won’t be able to get everything from one of these worldwide clouds and because of special laws within your own country we’ll still work with a regional cloud provider. Who will these players be? I believe the Telco’s will own these clouds much like they own the telephony infrastructure today. I also believe we already see this trend taking shape. Verizon acquired Terremark, Time Warner acquired Navisite, Windstream acquired Hosted Solutions and Cbeyond acquired MaximumASP. So we’re seeing Telcos and CLEC’s acquiring hosting companies already.
Does that mean that companies like Applied Innovations have a shelf life then? Absolutely not. You see, I believe companies like mine will have to leverage infrastructure components from both worldwide and regional providers. But there will be specific business vertical needs that won’t be provided from these larger providers and that’s where a niche focused company like Applied Innovations comes in. We’ll be able to offer services on top of these other infrastructures and that will suffice for about 80% of the market, but that other 20% of the market (yeah, I’m falling back on the 80/20 rule) will not be served by that infrastructure and they’ll be prime candidates for upselling to a niche focused cloud infrastructure and that’s where companies like mine will come into play. We’ll offer premium services focused on specific business/vertical needs.
#3 What Happens To RackSpace and SoftLayer?
It’s about this point in the conversation that someone asks “Well, what about RackSpace? Will they be one of those 5?” and I believe they will try to be but I don’t know if they will or will not. You see, I think Rackspace would be an ideal acquisition for AT&T and you’ll note Time Warner and Verizon have made their acquisitions already but AT&T hasn’t. So the question is more about who will AT&T acquire? I don’t think they’ll acquire Rackspace and from what I’ve seen, I think Rackspace’s ego is the reason they won’t be acquired. They were for years after all, the company every looked to as the giant in our industry. Only while everyone was so focused Microsoft and Google and worried about how to fight against them, this book company (Amazon) comes along and quickly takes over that position as the largest cloud provider. Make no mistake Mr. Hosting Company, Amazon has eaten your breakfast and they’re eyeing your lunch.
You see, I think Rackspace is large enough that they could be one of those worldwide clouds, the only thing I don’t know is if they would get acquired or not. The company I see that would be an ideal fit there would be AT&T but I’m betting that conversation has probably already happened and didn’t fair well. The reason I’m betting that is that it seems AT&T’ is using SoftLayer to power their SMB offering. If you look at the previous Telco purchases of hosting companies: Time Warner acquired Navisite, who powered their SMB offering, Verizon acquired Terremark, who powered their SMB offering and AT&T? Who powers AT&T’s SMB offering? It seems it’s not Rackspace but rather SoftLayer! If that’s the case then I’d put my money on SoftLayer eventually being acquired by AT&T and thus it seems Rackspace has to be one of those global clouds if they don’t want to go the way of Kodak.
#4 How’s OpenStack Play Into This?
Prior to starting AppliedI, I was an electrical engineer for Motorola. I had the privilege to work in a group that was building a next generation paging protocol. Prior to this protocol all of paging worked over an analog infrastructure and the new protocol was a digital paging protocol.
I watched our senior engineers (and these are the guys that had dozens of patents and fellowship status within Motorola) work their butts off to establish this protocol they developed as the industry standard protocol and I asked our VP why was the protocol so important? I mean we were making pagers (beepers was kind of a dirty word) and what did it matter who made the protocol, we were just making hardware. What I was told was that it was important Motorola own the standard. That if our protocol became the standard then we could define the direction the rest of the industry takes and we’d have a unique competitive advantage as we were already working on radios (pagers) that once the standard was defined we could market to our customers, months (if not years) ahead of our competitors! This protocol became known as FLEX and did become the standard and for a brief period (while cellular was too expensive) it won but over time cellular and SMS won over pagers. I’m also pleased to say I had a role in that happening (a very very small role but a role nonetheless).
I think this same idea of becoming the industry standard is why solutions like OpenStack, CloudStack, Hyper-V and VMware are all so important. I think Rackspace’s OpenStack is their play to become the cloud stack of choice and if they win that, they’ll win that same advantage that Motorola won with FLEX!
It’s no secret Microsoft and Rackspace have a close relationship but I often wonder just how close these days? I doubt we’ll see Rackspace touting Azure or Hyper-V over OpenStack and if they do then we can predict that OpenStack will fail. So the coming months will be interesting to see what happens.
#5 Who Will Win The Cloud Stack Standard?
CloudStack, Citrix’s play for cloud standard was born from their acquisition of cloud.com. I believe CloudStack is another contender for this title and with it’s Amazon EC2 compatible API you’d think it would be the horse Amazon would back but recently Amazon started backing Eucalyptus so who knows what will happen there. I think CloudStack does the best job of being Hypervisor agnostic though.
Hyper-V, make no mistake, Hyper-V will succeed in owning the lion’s share of the on premise cloud business. I believe whoever owns the on-premise business will also own the lion’s share of the public cloud and thus I believe that will be Microsoft. Unless they too want to go the way of Kodak! One could argue that they are headed in that direction already today and Azure, Hyper-V and more importantly System Center are their ticket out of that path. But it’s no secret that Microsoft is a large company composed of a number of individuals with competing egos so this will be interesting.
VMware, I’m not sure what will happen to VMware. I don’t think they’ll go away and I wonder if they will be acquired away from EMC by someone else. What is interesting is that VMware for so long made their money on the Hypervisor and now the hypervisor is pretty much free (we can thank open-source for this, so thank you!). I believe VMware raised their prices because they want to milk the cow. They want to get as much money as they can from their hypervisor business for as long as they can and they’re taking that money and putting it today R&D on what will be the future for VMware. One thing is certain, Paul Maritz is an amazingly bright individual and extremely capable CEO so I wouldn’t count VMware out.
Parallels, I’d be remiss if I didn’t discuss Parallels as they have helped AppliedI’s growth tremendously. Unfortunately, I think Parallels needs to stop focusing on their own hypervisor and trying to own the entire stack for hosting providers. I think OnApp has come in and is giving them a run for their money today and I think Parallels needs to quickly become hypervisor agnostic and could continue to win this niche. What’s advantageous for Parallels is that they own the hosting automation market today. If we equate them to how cloud will play out then Parallels could adopt these other hypervisors or even cloudstacks (be it cloudstack or openstack) sit on top of that and offer all of the features and functionality of those solutions with their hypervisor of choice. THEN, when you reach limitations of those solutions all Parallels would have to do is pitch you on why their proprietary solution is better and win the easy upsell. So Parallels is uniquely positioned here and an adoption of openstack by Parallels wouldn’t be a bad move at all in my opinion and in fact could open up opportunities for them in enterprise markets that they aren’t in today!
Those are my 5 cloud predictions and are truly just the view of the game from where I sit. I think you can read through that though and get a pretty good idea of the direction I believe Applied Innovations has to navigate and the direction we’ll head. I’d love some feedback and comments and curious if you think there’s something I’m overlooking.