If you've been continuing to run your IT infrastructure as-is, without keeping up on industry trends -- look out, because there's a storm coming that's going to take you by surprise.
This storm will factor into your decision making process as you consider your future IT spend and whether you continue to keep your workloads entirely on-premises, in a hybrid configuration, or in a hosted Private or Public Cloud.
While there are a number of innovations related to the cloud that are all seemingly occurring at once that are accelerating Cloud adoption, there are three in particular that will heavily impact your existing TCO on your IT infrastructure. I call these the 3 "Cs" of IT disruption.
The first "C" is Containers
As I explained in detail recently about what Containers are and what the overall benefits to your organization can be for implementing them, this technology alone will vastly increase the density of your virtualization hosts on-premises and also at hyperscale cloud providers, be they public or private.
Here's the bottom line: an increase in workload density through the use of container technology means higher levels of efficiency as well as the ability for you to reduce internal infrastructure costs, as well as for a cloud provider to reduce their own internal infrastructure costs and then pass that savings down in the form of cheaper IaaS and PaaS offerings to you as a tenant/subscriber.
Given the greater economies of scale that hosters or hyperscale providers have over your own internal IT infrastructure, this means you are going to have to (yet again) adjust your calculus for what workloads still make sense for you to keep on-prem.
While Containerization is primarily an Open Source and Linux-based technology today, we're not far off from it being a Windows and Microsoft-based technology as well, given current developments in Windows Server 2016 and ongoing developments at Azure.
So while this may not necessarily be a technology you are going to easily adopt within your own datacenter within in the next six months, it definitely should be on your roadmap within the next year or two.
At the bare minimum, you'll likely see the hyperscalers reduce their IaaS, PaaS and SaaS pricing first, which may give you enough incentive to lighten your own internal burden and also reconsider your ongoing software licensing strategy.
The second "C" is Convergence
Specifically I'm referring to converged network infrastructure, using 10Gig and faster Ethernet fabric for both your application and storage traffic.
If you (or your cloud provider) is using Microsoft Cloud Platform and Azure Stack, the use of NVGRE virtual networking technology will eliminate the 4096 physical VLAN barrier on your switch infrastructure and also eliminate the need for a separate fabric for your storage, such as Fiber Channel.
Along with NVGRE, tighter control over that converged network using other vendor-specific technologies such as Cisco ACI software defined networking (SDN) will also reduce your (and your provider's) costs, regardless of what hypervisor or cloud stack they use.
In the case of ACI, you or your provider will need to upgrade to a more modernized switch infrastructure (such as Cisco's 9000 series) but that upgrade will permit much higher levels of capacity as well as the ability to provide superior quality of service.
With technologies such as NVGRE and ACI, the "Noisy Neighbor" phenomenon within public clouds will be a thing of the past, and cloud SLA's will align much more closely to what you are used to withn your own private datacenter infrastructure.
The third "C" is Commoditization
While Commoditization covers a broad spectrum of technologies that can be installed in your own or provider's datacenters, whether it be denser and cheaper servers or any number of other things (comprehensive single-vendor cloud software stacks that offer cheaper, all-inclusive licensing is another example) but the one that I think will immediately affect the private datacenter and the cloud provider will be Commodity Storage, specifically the proliferation of JBODs over traditional storage arrays.
JBODs have been around for a long time, but their adoption is going to be accelerated by the previous "C" on this list, Converged Networks. With the use of the SMB 3 protocol and InfiniBand-compatible switches/10Gbps RDMA NICs on your virtualization hosts and file server heads, you'll be able to get SAN (or considerably better than SAN) speeds over converged Ethernet.
Additionally, due to the converged network, you (and of course, your cloud provider) will be able to do high-availability between JBODs using tiers of inexpensive SATA SSD and spindled drives rather than pricy SAS drives and clustered interconnects, using network-based replication between file server heads.
What does that mean overall? Cheaper cloud storage, cheaper HA and cheaper disaster recovery solutions.
Are the 3 "Cs" a perfect storm for Cloud adoption? Talk Back and Let Me Know.
See related coverage:
- Here comes the future of application development: Treating infrastructure as code
- Containers are a developer's dreamscape
- Containers: Fundamental to the cloud's evolution