Head in the cloud: 3Tera CEO defends cloud computing
- TAGS:3tera, Barry X Lynn, cloud computing
- IT TOPICS:Emerging Technology, Enterprise Software & Services, Internet, Servers & Data Center, Storage
I talked to Barry X Lynn -- the Chairman and CEO of the cloud-computing company 3Tera -- about moving to a cloud architecture. Is he the Forrest Gump of IT? Read on to find out.
JB: Large companies worry about cloud security -- why should they not be too concerned? And, what about privacy issues -- who owns the data, and what if it gets lost?
BXL: Information technology assets are more secure in a cloud than they are in today's traditional "physical" data centers. I'll say that again - IT assets are safer in a cloud than they are in a traditional data center.
Of course, this is only true of clouds done right. cloud computing is great, wonderful, magical and incredible, but it is no substitute for bad architecture.
In a traditional data center, moving an application of any complexity to another data center is a bear. But when the entire application is encapsulated with all the resources it requires, and is completely abstracted from its hardware, migrating it becomes no more difficult than migrating a simple file or object. So, if you have an application running in a data center whose security becomes dubious, you have the option to easily migrate that application to a more secure data center.
Let's face it: even if you are a giant financial institution, pharmaceutical company, etc., you are not upgrading your data centers every time new technologies leapfrog the old. You can't. These days, that's happening almost monthly.
But, if you are in the data center operations business, you can't afford not to upgrade everything, including your security technology, frequently. If you don't have stuff that's as good as the other guys, you'll be out of business. So, the ability to move applications from your data centers to those of professional data center operators, and the ability to easily move around among these most secure data centers, affords you the luxury of choosing to always run your applications in the most secure places available.
More importantly, the nature of cloud computing makes applications more immune from hackers and evil doers. Applications can be configured to migrate naturally among multiple data centers as part of the course of their operation. Imagine the ability to have your applications run in different data centers arbitrarily each time they are fired up - or, better yet - having them move around periodically while running. You will have turned your applications from sitting ducks into moving targets. Serious hackers out to cause harm and commit sabotage will want to monitor applications for a period of time to learn about their vulnerabilities and the best times to insert spurious, fraudulent transactions and data. How will they do that when they don't know where the heck they are?
With regard to who owns the data and what if it gets lost: The database owns the data. If it gets lost, and the application is mission critical, you are up the creek without a paddle. That's why you want your data in a well architected cloud with inherent redundancy. In short, my answer has to be, don't lose your data. Cloud computing done right assures that.
JB: Is the cloud more for companies that do marketing or engineering but not banks or health providers?
BXL: No, not at all. Cloud computing is a horizontal technology. It is, and should be, agnostic with regard to the vertical applications running in it.
Let's be clear. Cloud computing is the next generation of what a data center is today. Cloud computing creates economies for small operations as it gives them access to world class technologies they ordinarily cannot afford, initially in small bite sized pieces and only pay for what they consume - no more. Cloud computing creates economies for huge organizations by precluding the need for gross over-provisioning in order to make sure resources are available in the rarest of high volume instances. Servers are more densely populated resulting in huge capital expense reduction. Administrators manage applications rather than servers resulting in huge operational expense reduction. Better yet, the dynamic provisioning of cloud computing, along with the simplicity of configuration and the abstraction of the applications from the physical infrastructure, shrinks times to market by an order of magnitude, resulting in great market and revenue opportunities
Finally, inherent in cloud computing is comprehensive disaster recovery. The abstraction of applications from their hardware turns applications and all of the resources they need to run into simple data that can be copied, stored, restored moved, and migrated just as easy as any other data.
Better security; more privacy; world class technology readily available; no need for over-provisioning; reduced capex and opex; faster time to market; increased market opportunities; increased revenues; comprehensive disaster recovery: well, I'd say cloud computing is for everyone.
JB: What year do you think all computing will be in the cloud? Why then?
BXL: It's impossible to say exactly when, but it's soon, and not soon enough.
With regard to web applications, the train has already left the station and is accumulating speed rapidly. With regard to enterprise data centers, the need to turn data centers into metered utilities, to get higher utilization out of servers, to reduce administration costs, to attain much more efficient deployment of applications, etc., are all problems screaming to be solved. Most data center operators have started some form of project to rationalize their data centers of the future and those plans include cloud computing. Once the efficiency of Web applications in the cloud is apparent - heck, it's already as plain as the noses on our faces - internal applications will be deployed in the cloud. Cloud storage technologies will continue to mature and will become the place of choice to deposit information that must be available any time from anywhere. Exactly how long this takes will depend on the willingness of IT folks to adopt innovation, the economy, competitive environments, globalization, sun spots, voodoo, etc. But these factors aside, it is already on its way. Information technology is the next great utility and it is inconceivable that its domain will remain millions of homogenous silos.
JB: Cloud dissenters like Richard Stallman say local, client computing is the only way to keep big brother away, what would you say to the opponents of the cloud?
BXL: Stallman says, simply, that cloud computing is "stupid". So, I guess I'll just have to claim to be the Forest Gump of information technology. He says client computing is the only way to keep big brother away? Client computing has historically been the most pervasive of hackers' targets.
The term "cloud computing" may or may not, in fact, be hype. What we are talking about is massively scalable, both up and down, information technology on demand. It's hard to conceive of that being "stupid". It's no more stupid than picking up a telephone, getting a dial tone and completing a call anywhere in the world through interconnected "clouds" of phone company networks, only paying for the infrastructure used to complete that particular call.
I will ponder this more as I pedal my bicycle as hard as I can to generate the electricity I need to operate my two watt bulb so I can see what I am doing as I string a paper cup to another one in my geriatric mother's nursing home so I can occasionally ask her how she's doing.
JB: Others say bandwidth will never get there, because it still fails, runs slow, and is not secure. Is that valid?
BXL: Bandwidth is already there and is increasing all the time. Saying that cloud computing won't work because networks can fail and slow down, is, frankly, preposterous. This is akin to saying automobiles can't work because engines can seize and gaskets can leak.
Regarding the claim that cloud computing will not work because networks are not secure - well, I don't see governments about to go to war issuing DEFCON orders using carrier pigeons instead of networks. I don't see financial institutions transferring trillions of dollars a day by packing cash into messengers' sacks rather than using the Fedwire. All businesses and government agencies use secure networks today, and have been for quite some time, as the safest, most private way to move data from one place to another. How does cloud computing change this? Secure networks are secure. Insecure networks are not. If transactions in clouds get compromised because of a network break in, that's a result of poor network security and not inherent in the cloud.
Cloud computing is the most disruptive information technology to come along in a very long time. The number of excuses that emerge to resist any change is directly proportional to the amount of disruption incurred by the change.
JB: What are some of the steps required for better cloud security going forward?
A major step forward will be the acceptance that cloud computing is global. "The Cloud" will be located in many geographically diverse locations. A major criterion for choosing locations for parts of "The Cloud" must be the security of those locations. Cloud architectures must be designed with built in redundancy. The nature of the cloud must be such that failing over, duplication of everything, abstraction to the point that any hardware configuration can run any applications is as much a part of the architecture as the operating systems, middleware, and databases that run in it.



