Entering the production era of cloud computing
The following is an excerpt from the recently published SolidFire whitepaper, Beyond Test & Development: What virtualization can teach us about the next phase of cloud computing. This is the third post in a six part series that will be published each Monday. The whitepaper in its entirety can be found here.
Seven years into cloud market adoption, and Amazon Web Services finds itself at a similar position to VMware after the first seven years in the x86 virtualization market. Projecting out the next phase of growth in cloud computing will be dependent on the ability for cloud service providers to architect a VMware-like market transition from primarily test and development workloads to production applications. However, no single service provider-not even AWS-can bridge this transition alone. The commitment and innovation across the cloud ecosystem must at least parallel that demonstrated by the ISV and hardware vendor community in support of VMware.
Vendors are already stepping up to support the cloud’s advancement by building in features to support multi-tenancy and self-service via automation, and leveraging new architectures to drive increased scale. Similarly, Amazon continues to add services to support production environments, including Route 53, VPCs, Direct Connect, and traditional relational database certifications. If the revenue growth realized by VMware from successfully navigating their transition is any indication, the potential returns to be had from investing in the production era of cloud computing will more than offset the upfront investments.
Figure 4 plots VMware’s actual license revenue growth over the ten-year period from ’02 (YR1) to ’12 (YR11). This data effectively captures both the test and development and production phases of the server virtualization market. Overlaid against this data set, with time-scale normalized at Year Zero, is the early cloud computing growth trajectory in its first seven years (using AWS S3 objects as the proxy).
Missing from this graph, as the history is still to be written, is what happens in the next phase of cloud computing. AWS recently announced that, only one-third of the way through 2013, the total of S3 objects stored had eclipsed the two trillion mark. If this pace continues, cloud adoption will show a significantly accelerated growth trajectory relative to the same period of x86 virtualization adoption by the end of 2013. Against this historical context, cloud computing’s initial adoption trajectory has been very impressive, especially considering we are now just scratching the surface of the opportunity for production applications.
Where to from here?
If the adoption profile witnessed in the x86 virtualization era is even remotely close to what we are now seeing in the cloud market, then the production era of cloud is upon us. The implication here is that real market growth in the future will be dependent upon the continued shift of both green-field and legacy production applications away from infrastructure silos of the past to the shared multi-tenant infrastructure profile that now defines cloud. However, to run properly in a cloud infrastructure, these production applications come with an entirely different set of infrastructure dependencies and service-level expectations compared to test and development and backup/archiving workloads. Concerns regarding application availability, performance variability, and scale all take on a completely different complexion when the discussion revolves around production applications.
In our next post in the series, slated for July 1st, we will touch on some of the different challenges faced by service providers as they navigate the transition to the production era of cloud computing.
-Dave Cahill, Director of Strategic Alliances
| 0 comments
Posted in Other.