Written by Roger Strukhoff, Director of Research Altoros - Conference Chair Cloud Expo
Private cloud seems to be stuck in neutral. Why? The answer may partially reside in the lack of an effective way to share very large datasets and files, a sort-of “last-mile problem” facing enterprise IT.
Here's Some Background
Public-cloud provider Amazon is now reaching toward $12 billion in annual revenue, a number that Microsoft will soon be claiming for its competing Azure public-cloud services.
But we've seen scant progress in private cloud. VMware, one of the companies that kicked off our present cloud era, has an annual run rate of about $6 billion. Yet its products are used to virtualize resources, not necessarily achieve key benefits of cloud such as usage metering, easy flexibility, and the illusion of infinite scalability.
Meanwhile, IBM claims success with its Bluemix strategy, which is based on Cloud Foundry, a platform originally designed for private cloud that spun out of VMware a few years ago.
But within the global $2.2 trillion enterprise IT market (a number from Gartner that does not include telco), we can surmise that private cloud has captured only a precious few percentage points of this budget.
What's the Delay?
As Conference Chair of Cloud Expo (which we hold every year in New York in June and Santa Clara in November), I'm privileged to witness the gamut of cloud software, hardware, and services at the events. I also read over hundreds of speaker proposals annually.
Yet ten years into the Cloud Computing era, only 5% of workloads and 9 - 10% of enterprise IT budgets have migrated to the cloud in all its forms, according to surveys we've done at Cloud Expo.
So what is impeding private cloud growth?
It could be many things. I've heard, for example, many arguments in recent months about enterprises and their migrations to cloud – discussions of buy vs. build, buy and build, all in to public, all in to on-premises, hybrid cloud, multi-cloud, multi-hybrid cloud, multi-datacenter-hybrid cloud, the role of mobility, and the ominously increasing emphasis on the Internet of Things.
There could easily be a lot of “analysis paralysis” going on, as executives seriously and slowly weigh alternatives that will affect their enterprises for years to come.
But they need to pick up the pace. Our current dynamic era is experiencing data growth of 26% per year, according to research by Cisco. More than a zettabyte (one million petabytes) of data will cross the Internet this year, with several more zettabytes being handled by datacenters internally.
At current growth rates, the amount of data created annually will double every 2.7 years, and reach a yottabyte (one thousand zettabytes) around the year 2040.
As I was thinking all this through, I realized that maybe it's not a case of analysis paralysis, but a simpler case of a technology bottleneck somewhere along the line.
In fact, I think there is a“last mile problem” at least partly responsible for the paradox of fast data growth but slow cloud adoption.
Because after all these years, there are still of dollars of enterprise applications immobilized in legacy silos and cannot consistently and effectively be shared across organizations.
Additionally, a large percentage of this data is contained in very large files. And today, organizations need to have this data available to people distributed across a region, country, or the globe. Visual data needs to appear the same way across all desktop and laptop systems, and numerous mobile devices.
But transmitting (ie, uploading and downloading) these files to local and remote systems and devices requires hours and even days, so is not practical. Rewriting applications from scratch is also time-consuming, expensive, and impractical.
Is VDI the Answer?
One poor solution has emerged: VDI, or virtual desktop infrastructure. This method involves “screen scraping” images and replicating them on other devices. The screen scraping keeps file sizes down, but does not account for different screen sizes and formats.
As the only way to go, VDI has been able to emerge as a $6 billion market. Yet its limitations are holding back what should be a massive migration of enterprise IT to multiple cloud architectures.
An effective way of making visual data easily accessible and interactive throughout the organization will spur a new period of enterprise IT migrating to cloud computing.
How About This?
Then I discovered a piece of enterprise software called PureWeb (www.PureWeb.io), developed by Calgary Scientific, headquartered in (you guessed it) Calgary, Alberta, Canada.
As a patented, API-driven platform, PureWeb allows original files to be rendered in their full quality and displayed faithfully on any system or device, through HTML 5, Android, or Apple IoS. (Note: This stuff is fascinating to me, and I'm not being compensated by Calgary Scientific to write this.)
By native rendering of full, original files to all systems and devices, PureWeb users can also interact with the images. Yet the files themselves are not transmitted, so security is maintained. No one but the sender has access to the data.
Now, visual apps that are trapped in legacy systems and poorly rendered through VDI can be liberated across the enterprise at a cost less than VDI and far less than re-writing existing apps. With the visualization bottleneck removed and last-mile problem solved, legacy applications can be on-boarded to the cloud in days or weeks instead of years.
The result is, in effect, the creation of new, "omni-channel" applications that are fully viewable and interactive on all systems and devices. This new reality in the cloud extends their lives and the investments companies have made in them.
Additionally, new apps can be developed with enterprise-wide rendering capability built in.
PureWeb might be that breakthrough technology that not only offers a superior solution in the use of large, visual files, but will also spur a new era of private-cloud growth. If enterprises are no longer stymied by the issue of faithfully rendering, sharing, interacting with, and securing their precious resources, they can then consider more effective ways of deploying their enterprise IT overall.
Solve the bottleneck, get rid of the last-mile problem, and suddenly you are free to migrate your internal IT resources to a 21st-century cloud architecture