Source: Gartner (December 2011)
Cloud Services Market Expanding to $109 Billion in 2012
It
wasn't too terribly long ago that "cloud computing" was a loosey-goosey
marketing term being thrown around by anyone and everyone in the
software space. And now? There's been a marked shift towards cloud-based
services, which is a market that research firm Gartner predicts will grow 19.6 percent to $109 billion by the end of 2012."The cloud services market is clearly a high-growth sector within the overall IT marketplace," said Ed Anderson, research director at Gartner. "The key to taking advantage of this growth will be understanding the nuances of the opportunity within service segments and geographic regions, and then prioritizing investments in line with the opportunities."
In 2011, Gartner says the total public cloud services market was $91.4 billion. By 2016, the firm forecasts it will top $206.6 billion.
Gartner's forecast is anything but absurd. Mobile users are increasingly dependent on the cloud, in part because they're somewhat forced to be, and also because of the convenience factor. Think about the number of media tablets that skimp on built-in storage, some of which don't offer a way to expand via a microSD card slot.
On the IT side, Gartner notes that business process as a service (BPaaS) is the largest cloud-based segment, accounting for more than two thirds of the market. In other words, firms are increasingly outsourcing their payroll, printing, ecommerce, and other business processes to the cloud.
While on the topic of cloud computing, be sure to check out our fancy cloud managing guide for some indispensable tips!
Will OpenStack Usher in a Cloud Revolution?
OpenStack is seen by many as the project that will, finally, make the cloud enterprise ready. Detractors don’t believe it and point to political in-fighting, domination by large vendors and a lack of maturity.
By Jeff Vance
If you’re already tired of all the hype surrounding cloud computing, you’d better brace yourself for another cycle focused on OpenStack, the open-source cloud platform that’s touted as the operating system for the cloud.
The forces aligning behind OpenStack are impressive ones. The project originated in NASA, was moved along in partnership with Rackspace and is now spearheaded by the OpenStack Foundation. OpenStack is backed by 180 member companies, including biggies like AT&T, HP, IBM, Cisco, VMware and Intel.
Oh, and the Foundation has $10 million in funding. Not bad.
Despite its roster of supporters, OpenStack has plenty of detractors too. Before acquiring network virtualization startup Nicira, recently added member VMware bad-mouthed OpenStack as an “ugly sister” to vCloud (the other ugly sisters being Citrix-led CloudStack and Eucalyptus).
I should also note that while you don’t hear much out of Amazon, AWS is the 800-pound, thus far silent, gorilla in the room.
Just when you were hoping the enterprise cloud picture was getting clearer, along comes political in-fighting about whose cloud is more open and which will meet the performance standards needed for enterprise-class cloud computing.
VMware Joins – Consensus or Kiss of Death
OpenStack was originally seen as an open alternative to VMware’s proprietary dominance over data center virtualization and what would eventually turn into proprietary clouds.
Boris Renski, EVP of cloud startup Mirantis and a Gold Member of OpenStack, isn’t happy about VMware’s participation in OpenStack. On the Mirantis blog, Renski wrote:
"Subduing OpenStack is exactly what VMware did by joining the foundation. Every enterprise considering OpenStack that we ever encountered at Mirantis was primarily interested in OpenStack as an open alternative to proprietary VMware. While in reality OpenStack and VMware are different kinds of beasts, perception-wise there is no argument: enterprises see OpenStack as a substitute for VMware. Now, with VMware in the OpenStack foundation, every enterprise buyer will rightfully ask the question: 'If OpenStack is not competing with VMware, then what the hell is OpenStack?'”
Not everyone in OpenStack feels this way. “Boris was one of two board members to vote against VMware. Two out of twenty-four,” said Josh McKenty, a co-founder of OpenStack and the CEO of Piston Cloud Computing.
“Look at hypervisors, they [VMware] have 90 percent market share. Every cloud has a hypervisor in it somewhere. VMware is in best position to make all of that work, and when we talk to end users, they all tell us they want VMware in OpenStack,” he said.
McKenty added that what is good for the overall OpenStack community in this case might not work out as well for Mirantis. We’ll see.
Performance and Maturity
Other detractors point to the relative immaturity of OpenStack and performance issues.
German company Dolphin IT Services downloaded OpenStack, gave it a test run, but then abandoned it.
“It became apparent quite fast that the product was not mature enough to be deployed productively. A lot needed to be done on our side in order for it to work correctly. There were still some features we dearly needed that were not implemented yet – or not sufficiently implemented – like billing and an appealing Web front end,” said Andreas Kunter, CEO of Dolphin IT Services.
The company instead adopted the cloud platform from startup OnApp. “[With OnApp] we could basically deploy out-of-the-box. It integrated nicely with our chosen billing platform, and allowed us to grow our business without having to pay large amounts upfront,” he said.
One of the main concerns Kunter had with OpenStack is support. “Support after deployment is often underestimated,” he said. “We know a lot about virtualization and hypervisor technology, but we are still learning while maintaining our cloud. When things go wild and you need to solve it fast, it is good to have proficient support to back you up.”
Plenty of companies, Rackspace included, intend to make money supporting OpenStack. However, it will take time to build those support teams out, and to get them up to speed on all of the ins and outs of the sprawling OpenStack project.
The other two major complaints that detractors have about OpenStack are: 1) big vendors like Cisco, HP, IBM, Intel and VMware could dominate the project in ways that won’t be beneficial to the larger community and 2) OpenStack performance isn’t yet competitive with the likes of AWS.
As for complaint one, anytime you have that many big vendors working together, people will be wary of them. Would it be better for them to simply offer an array of competing products, rather than putting their proprietary layers and services over a common core?
At the very least, the threat of vendor-lock, a very real cloud concern, is lessened with OpenStack.
Complaint two is trickier.
Cloud storage startup Nasuni has tested all the major bulk cloud storage service providers for performance, stability and scalability. “We’re neutral about OpenStack,” said Andres Rodriguez, CEO of Nasuni. “But we’re going to rigorously test any product or service we may offer to our own customers.”
Rodriguez pointed out that Nasuni didn’t specifically test OpenStack, but since Rackspace developed and uses OpenStack, Nasuni regards Rackspace Cloud as the OpenStack showcase. Based on benchmarks Nasuni set, only six out of the sixteen cloud storage providers they tested passed.
Rackspace did indeed pass, but the cloud storage test found that Rackspace lagged far behind leader Amazon in a number of key metrics, including speed, both read and write errors, and stability.
“Remember, cloud storage is all about scale, and there’s a huge gap between Amazon and everyone else, just based on scale alone. Most of Rackspace’s business is still colocation. Their cloud storage footprint is dwarfed by Amazon,” he said.
While cloud storage is new to most vendors, Amazon has been working away on this problem for the past twelve or fifteen years. “By the time they went public, they’d already accumulated seven years or so of real-world operational experience. Challengers are coming in a decade late,” he added.
The Future of OpenStack
Storage is only one part of OpenStack, and arguably one of its least mature parts. It’s premature to write it off as something that will never challenge Amazon, especially as storage demands continue to skyrocket.
And not all storage is the same. Mission-critical storage is much different than, say, backups of non-critical media files.
McKenty, the OpenStack co-founder, said that his next major step for OpenStack is to get the separate projects better joined beneath a common framework. With so many developers working independently, the projects can grow apart. “It’s a challenge to figure out how to keep the projects loosely coupled, so you don’t stifle the creativity of developers, yet linked, so everything works well together,” he said.
He believes it’ll be another release or two before they reach that goal, but when they do, OpenStack will have a single command line that will ensure interoperability at the most basic level.
“We’ve learned plenty of lessons along the way,” said Wayne Walls, one of Rackspace’s key OpenStack developers. “We quickly learned that you could have about a billion different variations of an OpenStack cloud where you tell people to download it and go run it as they see fit. But that’s very hard to support at scale.”
Walls believes that OpenStack is nearly past its awkward “immature” phase. Companies are taking OpenStack and developing products around it. If you look at some of the startups in the OpenStack Foundation, including Mirantis and Piston Cloud Computing, all of their messaging focuses on OpenStack. You would have a hard time getting much VC funding if OpenStack was a doomed science project.
“OpenStack has 550,000 lines of code today, with close to 500 developers around world contributing code,” Walls said. “This model pushes best of breed to the top. Whether it’s networking or storage or whatever else, the world’s top experts decide how things are done.”
The other key variable that Walls pointed to is the fact that no matter “how deep in the weeds” you get with OpenStack, it’s very easy to pick up your assets and move them. It’s not nearly as easy to do so with closed APIs.
Of course, this is important if enterprises are ever going to gain confidence in public clouds, but it’s equally important for private clouds. This kind of portability means an OpenStack cloud, almost by definition, is a hybrid cloud, allowing you to move back and forth between private and public environments almost at will.
That’s a competitive advantage that OpenStack is gaining that competing solutions will have trouble keeping up with.
By Jeff Vance
If you’re already tired of all the hype surrounding cloud computing, you’d better brace yourself for another cycle focused on OpenStack, the open-source cloud platform that’s touted as the operating system for the cloud.
The forces aligning behind OpenStack are impressive ones. The project originated in NASA, was moved along in partnership with Rackspace and is now spearheaded by the OpenStack Foundation. OpenStack is backed by 180 member companies, including biggies like AT&T, HP, IBM, Cisco, VMware and Intel.
Oh, and the Foundation has $10 million in funding. Not bad.
Despite its roster of supporters, OpenStack has plenty of detractors too. Before acquiring network virtualization startup Nicira, recently added member VMware bad-mouthed OpenStack as an “ugly sister” to vCloud (the other ugly sisters being Citrix-led CloudStack and Eucalyptus).
I should also note that while you don’t hear much out of Amazon, AWS is the 800-pound, thus far silent, gorilla in the room.
Just when you were hoping the enterprise cloud picture was getting clearer, along comes political in-fighting about whose cloud is more open and which will meet the performance standards needed for enterprise-class cloud computing.
VMware Joins – Consensus or Kiss of Death
OpenStack was originally seen as an open alternative to VMware’s proprietary dominance over data center virtualization and what would eventually turn into proprietary clouds.
Boris Renski, EVP of cloud startup Mirantis and a Gold Member of OpenStack, isn’t happy about VMware’s participation in OpenStack. On the Mirantis blog, Renski wrote:
"Subduing OpenStack is exactly what VMware did by joining the foundation. Every enterprise considering OpenStack that we ever encountered at Mirantis was primarily interested in OpenStack as an open alternative to proprietary VMware. While in reality OpenStack and VMware are different kinds of beasts, perception-wise there is no argument: enterprises see OpenStack as a substitute for VMware. Now, with VMware in the OpenStack foundation, every enterprise buyer will rightfully ask the question: 'If OpenStack is not competing with VMware, then what the hell is OpenStack?'”
Not everyone in OpenStack feels this way. “Boris was one of two board members to vote against VMware. Two out of twenty-four,” said Josh McKenty, a co-founder of OpenStack and the CEO of Piston Cloud Computing.
“Look at hypervisors, they [VMware] have 90 percent market share. Every cloud has a hypervisor in it somewhere. VMware is in best position to make all of that work, and when we talk to end users, they all tell us they want VMware in OpenStack,” he said.
McKenty added that what is good for the overall OpenStack community in this case might not work out as well for Mirantis. We’ll see.
Performance and Maturity
Other detractors point to the relative immaturity of OpenStack and performance issues.
German company Dolphin IT Services downloaded OpenStack, gave it a test run, but then abandoned it.
“It became apparent quite fast that the product was not mature enough to be deployed productively. A lot needed to be done on our side in order for it to work correctly. There were still some features we dearly needed that were not implemented yet – or not sufficiently implemented – like billing and an appealing Web front end,” said Andreas Kunter, CEO of Dolphin IT Services.
The company instead adopted the cloud platform from startup OnApp. “[With OnApp] we could basically deploy out-of-the-box. It integrated nicely with our chosen billing platform, and allowed us to grow our business without having to pay large amounts upfront,” he said.
One of the main concerns Kunter had with OpenStack is support. “Support after deployment is often underestimated,” he said. “We know a lot about virtualization and hypervisor technology, but we are still learning while maintaining our cloud. When things go wild and you need to solve it fast, it is good to have proficient support to back you up.”
Plenty of companies, Rackspace included, intend to make money supporting OpenStack. However, it will take time to build those support teams out, and to get them up to speed on all of the ins and outs of the sprawling OpenStack project.
The other two major complaints that detractors have about OpenStack are: 1) big vendors like Cisco, HP, IBM, Intel and VMware could dominate the project in ways that won’t be beneficial to the larger community and 2) OpenStack performance isn’t yet competitive with the likes of AWS.
As for complaint one, anytime you have that many big vendors working together, people will be wary of them. Would it be better for them to simply offer an array of competing products, rather than putting their proprietary layers and services over a common core?
At the very least, the threat of vendor-lock, a very real cloud concern, is lessened with OpenStack.
Complaint two is trickier.
Cloud storage startup Nasuni has tested all the major bulk cloud storage service providers for performance, stability and scalability. “We’re neutral about OpenStack,” said Andres Rodriguez, CEO of Nasuni. “But we’re going to rigorously test any product or service we may offer to our own customers.”
Rodriguez pointed out that Nasuni didn’t specifically test OpenStack, but since Rackspace developed and uses OpenStack, Nasuni regards Rackspace Cloud as the OpenStack showcase. Based on benchmarks Nasuni set, only six out of the sixteen cloud storage providers they tested passed.
Rackspace did indeed pass, but the cloud storage test found that Rackspace lagged far behind leader Amazon in a number of key metrics, including speed, both read and write errors, and stability.
“Remember, cloud storage is all about scale, and there’s a huge gap between Amazon and everyone else, just based on scale alone. Most of Rackspace’s business is still colocation. Their cloud storage footprint is dwarfed by Amazon,” he said.
While cloud storage is new to most vendors, Amazon has been working away on this problem for the past twelve or fifteen years. “By the time they went public, they’d already accumulated seven years or so of real-world operational experience. Challengers are coming in a decade late,” he added.
The Future of OpenStack
Storage is only one part of OpenStack, and arguably one of its least mature parts. It’s premature to write it off as something that will never challenge Amazon, especially as storage demands continue to skyrocket.
And not all storage is the same. Mission-critical storage is much different than, say, backups of non-critical media files.
McKenty, the OpenStack co-founder, said that his next major step for OpenStack is to get the separate projects better joined beneath a common framework. With so many developers working independently, the projects can grow apart. “It’s a challenge to figure out how to keep the projects loosely coupled, so you don’t stifle the creativity of developers, yet linked, so everything works well together,” he said.
He believes it’ll be another release or two before they reach that goal, but when they do, OpenStack will have a single command line that will ensure interoperability at the most basic level.
“We’ve learned plenty of lessons along the way,” said Wayne Walls, one of Rackspace’s key OpenStack developers. “We quickly learned that you could have about a billion different variations of an OpenStack cloud where you tell people to download it and go run it as they see fit. But that’s very hard to support at scale.”
Walls believes that OpenStack is nearly past its awkward “immature” phase. Companies are taking OpenStack and developing products around it. If you look at some of the startups in the OpenStack Foundation, including Mirantis and Piston Cloud Computing, all of their messaging focuses on OpenStack. You would have a hard time getting much VC funding if OpenStack was a doomed science project.
“OpenStack has 550,000 lines of code today, with close to 500 developers around world contributing code,” Walls said. “This model pushes best of breed to the top. Whether it’s networking or storage or whatever else, the world’s top experts decide how things are done.”
The other key variable that Walls pointed to is the fact that no matter “how deep in the weeds” you get with OpenStack, it’s very easy to pick up your assets and move them. It’s not nearly as easy to do so with closed APIs.
Of course, this is important if enterprises are ever going to gain confidence in public clouds, but it’s equally important for private clouds. This kind of portability means an OpenStack cloud, almost by definition, is a hybrid cloud, allowing you to move back and forth between private and public environments almost at will.
That’s a competitive advantage that OpenStack is gaining that competing solutions will have trouble keeping up with.
Why corporate strategy needs to change with the cloud
By Prabhakar Gopalan
Prabhakar Gopalan looks at how the skill sets and mindsets of
corporate strategists need to change in the cloud computing era. This
means everything from learning to accept new revenue models to being
technologically savvy enough to actually build new products.
Cloud computing changes everything, including corporate strategy as
a practice. I have listed five reasons why, although I’m sure there are
many more. Long story short: Corporate strategists need to get out of
their 20th century mindset and into the 21st century.
Cloud computing doesn’t operate in the intentional strategy space. There are a lot of unknowns, many of which can change rapidly. A small firm could develop something valuable very quickly, scale it to millions of users in a very short time and all equations about competitor reaction, supply and demand forecasts become irrelevant (hey, that’s why we have auto scaling!). The frameworks have to be discarded for more agile ways of solving problems.
Want strategists? Hire technologists and entrepreneurs who can talk real subject matter, produce prototypes and demos, and delight your customers. Who wants another boring PowerPoint that debates whether the next big widget market is going to be $5 billion or $500 billion next year? Hire makers, and go make it.
Traditional corporate strategy teams staffed with ex-big consulting firm consultants should find the exit door as soon as they can if they are in a meeting to discuss growth plans incorporating the cloud. Move them to areas of business strategy that don’t have a large impact due to cloud computing (are there still some?) or retrain them. Even the billionaire mayor of New York City wants to retrain himself and learn coding!
GrubHub CEO Matt Maloney was right on when he said, “The take-away was that a strategy may play out on paper, but the only way to truly test the validity of your product is to put it in front of customers.”
Firms
that are used to billion-dollar businesses built largely through
lock-in, slow incremental-feature-driven innovation after years of
squeezing the customers, and near monopoly in market segments are
finally seeing how the cloud’s production, distribution and pricing
channel are changing the economics of growth. A large software firm used
to getting thousands of dollars on licensed software now has to compete
with a nimble software firm that can deliver the same or better
software for just a few dollars a month. Production, distribution and
pricing have all changed the basis of competition.
To compete in this space, corporate strategy has to reset expectations of what growth means –can larger firms deal with the smaller doses of revenues? It took Amazon Web Services almost eight years to get to top a billion dollars in revenue (see How Big is AWS) while the closest competitor, Rackspace, has taken almost the same time to get to $188 million in annual revenue.
So, the road to large revenues and profits is going to be much longer and more measured. For example, the average revenue per user (ARPU) for Evernote, a popular SaaS note-taking application, is less than $2 for most versions of its product.
The problem is that the synergy component depends on cutting costs by a lot. And when you welcome the cost cutters, they don’t know what inspires and spurs innovation and experimentation. They go with their chopping block and cut costs across the board. R&D budgets get tossed out and what you get is an innovative cloud startup squeezed in a non-cloud firm that doesn’t understand the factors for generating continuous innovation.
The next is IP acquisition. Here again, the IP that many of the innovative firms have developed is in design — user interface and user experience — which means discarding the old and making something really simple and easy to use. It is hard to measure that IP in traditional terms. Further, most of the development is based on open source and crowd sourced technologies, so there’s not a whole lot of proprietary technology to own when you buy a cloud computing startup.
Lastly, the distribution channel of the incumbents is of very little use because cloud services aren’t delivered in the same way previous services are delivered. Thousands of firms are developing applications over weekend hackathons using open source technologies and uploading them to app stores or their own sites accessible to anyone with an Internet connection. Few incumbents have significant investments in the value chain of these developers.
Prabhakar Gopalan is an entrepreneur and a product guy. Opinions expressed here do not reflect those of his employers. You can follow Prabhakar on Twitter: @PGopalan. Prabhakar is the founder of Simple Idea Labs, and kanban2go.com is one of their first experiments. Prabhakar writes and talks about products, strategy and chaos.
1. Emergent strategy rules
For years, the practice of strategy has been about analyzing value chains, applying frameworks like Porter’s five forces or newer strategic-intent-driven ideas like Blue Ocean Strategy. The problem with those framework-driven ideas is they assume a very static, deterministic model of the world. They work when the variables required to solve a problem are already well known, few in number and change at a slow pace.Cloud computing doesn’t operate in the intentional strategy space. There are a lot of unknowns, many of which can change rapidly. A small firm could develop something valuable very quickly, scale it to millions of users in a very short time and all equations about competitor reaction, supply and demand forecasts become irrelevant (hey, that’s why we have auto scaling!). The frameworks have to be discarded for more agile ways of solving problems.
2. Subject matter expertise in technology matters
As traditional growth frameworks and models are rendered useless (see No. 1 above), so too are the operators of those models. There’s very little a consultant with an MBA and no subject matter expertise in the cloud or its underlying technologies can advise your firm on what it means to design or develop a product for the future.Want strategists? Hire technologists and entrepreneurs who can talk real subject matter, produce prototypes and demos, and delight your customers. Who wants another boring PowerPoint that debates whether the next big widget market is going to be $5 billion or $500 billion next year? Hire makers, and go make it.
Traditional corporate strategy teams staffed with ex-big consulting firm consultants should find the exit door as soon as they can if they are in a meeting to discuss growth plans incorporating the cloud. Move them to areas of business strategy that don’t have a large impact due to cloud computing (are there still some?) or retrain them. Even the billionaire mayor of New York City wants to retrain himself and learn coding!
3. Product teams inform corporate growth in the cloud, not corporate strategy teams
If you are reaching out to your corporate strategy team to figure out what the next wave of innovation is, you have already missed the boat. Ask your product strategy team that is in front of your customer every day. They can tell you where the next pocket of growth is going to be. They are spending the time with the customer and know what to make while the corporate drones are still analyzing the spreadsheets and profit pools of a business that is past its prime or analyzing a market entry problem when you are already locked out of the market.GrubHub CEO Matt Maloney was right on when he said, “The take-away was that a strategy may play out on paper, but the only way to truly test the validity of your product is to put it in front of customers.”
4. Long tail + subscription economy = Granular growth
Firms
that are used to billion-dollar businesses built largely through
lock-in, slow incremental-feature-driven innovation after years of
squeezing the customers, and near monopoly in market segments are
finally seeing how the cloud’s production, distribution and pricing
channel are changing the economics of growth. A large software firm used
to getting thousands of dollars on licensed software now has to compete
with a nimble software firm that can deliver the same or better
software for just a few dollars a month. Production, distribution and
pricing have all changed the basis of competition.To compete in this space, corporate strategy has to reset expectations of what growth means –can larger firms deal with the smaller doses of revenues? It took Amazon Web Services almost eight years to get to top a billion dollars in revenue (see How Big is AWS) while the closest competitor, Rackspace, has taken almost the same time to get to $188 million in annual revenue.
So, the road to large revenues and profits is going to be much longer and more measured. For example, the average revenue per user (ARPU) for Evernote, a popular SaaS note-taking application, is less than $2 for most versions of its product.
5. Revisiting growth via M&A
Large incumbents of non-cloud technologies have resorted to M&A as a growth path. The business cases for these M&A projects are based on synergy, acquiring intellectual property and distributing it through existing channels that were developed over many years.The problem is that the synergy component depends on cutting costs by a lot. And when you welcome the cost cutters, they don’t know what inspires and spurs innovation and experimentation. They go with their chopping block and cut costs across the board. R&D budgets get tossed out and what you get is an innovative cloud startup squeezed in a non-cloud firm that doesn’t understand the factors for generating continuous innovation.
The next is IP acquisition. Here again, the IP that many of the innovative firms have developed is in design — user interface and user experience — which means discarding the old and making something really simple and easy to use. It is hard to measure that IP in traditional terms. Further, most of the development is based on open source and crowd sourced technologies, so there’s not a whole lot of proprietary technology to own when you buy a cloud computing startup.
Lastly, the distribution channel of the incumbents is of very little use because cloud services aren’t delivered in the same way previous services are delivered. Thousands of firms are developing applications over weekend hackathons using open source technologies and uploading them to app stores or their own sites accessible to anyone with an Internet connection. Few incumbents have significant investments in the value chain of these developers.
What to look for in your strategy leadership?
Companies need strategic decision makers and executives with the ability to overcome their own cognitive biases against radical new methods of delivering services to customers. In other words, they need executives that view the world through the lens of modern computing and business paradigms, can take delight in looking at the latest app and getting curious about its AAARR metrics, and can take a startup approach to market size and traction. These things will significantly effect their ability to experiment and grow corporate strategy practice in the brave, new cloudy world.Prabhakar Gopalan is an entrepreneur and a product guy. Opinions expressed here do not reflect those of his employers. You can follow Prabhakar on Twitter: @PGopalan. Prabhakar is the founder of Simple Idea Labs, and kanban2go.com is one of their first experiments. Prabhakar writes and talks about products, strategy and chaos.
Subscribe to:
Posts (Atom)