April 2009 Archives

How would you like to put 100 DVDs on one CD?



This week the folks at G.E. announced their latest breakthrough in Holographic Storage on Brian Lawrence's blog: it's about " threshold recording behavior" - the team at GE has apparently figured out how to improve performance by 100x to 200x over past efforts. 

For some reason, the folks over at the Fusion Render Cloud think in terms of either/or, not together. In her article "GE's holographic disc challenges cloud streaming," journalist Sylvie Barak describes their reaction to this technical breakthrough from GE.  Jules Urbach, founder and CEO of Otoy and Lightstager grumbles: "Live streaming is the real next frontier, not bigger discs..."  His argument is that storage has to live in the cloud, and that "going back to storing everything on a fixed disc the consumer has to physically possess seems like a step backwards to me."

Local, physical storage continues to play a significant role. Over time, it may be that physical "permanent" storage will only be cloud based, but that is a security and retention issue, not a performance issue.  When and only when network speed combined with compression technology delivers the same performance as local storage can the cloud be a replacement for local storage.

What Cloud Storage does do is deliver additional benefits, versus replacing local storage.  Having local and cloud storage work together is where both deliver significant benefits.  Keep a local copy of most recently used files, but the authoritative source of the file is in the cloud, where it is securely stored and backed up, auto-magically!  As local files are changed, the new file is sent to the cloud.  If the file on the cloud is changed, in a remote access use case, the local file is updated.  Local files should be a cache only, and when you call for a file that is not in the local cache ( a cache miss), the small time to import from the cloud is OK, because of the benefits of the cloud as authoritative source and "home base" for your files.

The Cloud should also be home to those files which are independent of the performance issues associated with access via the Internet, or where cloud capabilities (sharing and collaboration, for example) outweigh the local performance desire.  Also, over time, applications will be designed to expect Internet access of files, and this will also drive cloud adoption.

Driving the ultimate storage cost down is desirable and unrelated to where it is stored.  Different technologies, including magnetic, optical and solid state, will be selected for the same reasons they always have, read/write characteristics, performance, cost, reliability, persistence, size, electrical footprint, etc.  This will march ever onward and is complementary to cloud usage. 

The more choices the better. 

With the Sun acquisition, Larry "what's a cloud?" Ellison has once again changed the game. Here are a few key points to think about:

1) Oracle becomes the end-to-end IT enabler - from apps to disks; that's the party line.

2) Oracle begins the journey to the Cloud, and begins to develop the end-to-end Enterprise Cloud experience.

3) Oracle embraces the open source movement by attacking  Microsoft with MySQL.

4) Oracle's gain is clearly SAP's loss. Exadata + Sun = the new business intelligence?

5) Oracle owns Java, period. Ellison described Java as "the single most important software asset we have ever acquired." BONUS: they get the JMX API thrown in with the deal, which allows them to monitor all manner of resources.

6) Oracle delivers Peoplesoft-as-a-Service or Seibel-as-a-Service with credibility. Maybe they won't buy Salesforce.

7) Oracle pushes Open Office as a cloud offering to further disrupt Microsoft.

8) Oracle makes Sun hardware profitable.

9) One stop shopping for all your IT, from Cloud to your own data center - is where we are headed. The period of détente is over - Cisco, HP, IBM, and Oracle are racing to go to end-to-end environments, which HP and IBM have proven as a viable business model. What happens to Dell?

While the rumors fly all over the cloudsphere, what's important in the days ahead is how Oracle chooses to embrace the cloud - will it be an open or closed embrace?

With IBM, we all knew it would have been an open cloud, with Oracle the story is not so clear at all.  The silence on Jonathan's blog is deafening.

And for those of us who said that Ellison was kidding about the cloud, let's remember who said "the network is the computer!"  Today's netbooks are cloud devices.

My prediction: Oracle becomes one of the "Big Four" for the Enterprise, and quickly changes it's tune on the Cloud.

What's next? Anybody think Microsoft/Dell is an interesting combination? 

The last few days have seen a lot of discussion over the McKinsey & Company report: Clearing the Air on Cloud Computing.   McKinsey said the cost of the Cloud is high when compared to a large enterprises ability to implement similar services in house.   They may, currently, be correct, but over time, this will likely be a differentiation that will erode.  Enterprises will need the Cloud, as new applications become Cloud dependent.  We have seen, for the first time, services move down the infrastructure stack.  For example, tagging, sharing and collaboration will become requirements for Cloud Storage.

McKinsey rightly stated that Public Cloud is the domain of the small and medium business.  What I have observed is that every year, larger companies join this view.  We now say it this way:  SMB and mid tier enterprise will turn to the Cloud.  Mid tier will continue to include larger and larger companies.

Also, each Enterprise is different, and in different financial circumstances.  Their existing IT infrastructure may be out of date, and require a very capital intensive retrofit to achieve the same economies as public cloud.  To the extent an Enterprise does not want to put it's capital to work in IT, in house computing will be a less popular option, regardless of opex costs.  The McKinsey view seemed very cost focused, and cost is driving Cloud consideration in the enterprise.

The key issue was raised, but did not receive in depth consideration!

Public Cloud providers must, I repeat, must build out mission critical service delivery capability, with security, management and SLAs that deliver the credibility and service level an enterprise will require to move their in house applications to the Cloud.  Its that simple, and that is the issue.  IT service providers get this, and will deliver it within their clouds.  Simple to say, a bit harder to do.  But, it will happen.

UPDATE
- Tim Bray lets McKinsey have it >>
- Amazon's James Hamilton has an insightful post here >>

I've also been invited to present at Dealmaker Media's Under the Radar (UTR) conference on April 24 in Mountain View, CA.

In this case I'll be talking about what makes Mezeo unique: the deployable Mezeo Cloud Storage Platform that is purpose built for the IT service provider market.

Register here >>

I'll be doing a presentation at Cloud Slam '09 titled:
"The Most Interesting Sub-plot in Cloud Computing is Cloud Storage"

Here's what I'll be talking about:

While much of the collective discussion has focused on cloud computing and trying to define it, cloud storage has not so quietly gone from concept to tangible business with real revenue opportunities. Those actively in the business speculate that .40 to .90 cents from every $1 spent on cloud computing is storage, and with the continued increase of data creation from applications and people, it's only going to go up. Giants like Amazon, Google, Microsoft and now Sun have thrown their hats into the cloud storage ring, but they are not the only players. The economics for trusted service providers like IT Hosting providers make it attractive for their entry into this space.

Cloud Slam '09 is the world's largest virtual cloud computing conference, so you can attend the conference no matter where you are.

See you in the cloud:
Tuesday, April 21, 2009 4:00 pm
Eastern Daylight Time (GMT -04:00, New York)

Nicholas Hoover informs us that General Electric is pilot testing the Cloud.  Here are a few key points:

• GE has tens of thousands of servers in 10 primary data centers, and the number of servers in those facilities is growing 20% annually.
• 60% of GE's servers are already virtualized.
• GE is already testing cloud platforms from three vendors: Cisco, Sun, and VMware
  
• CTO Greg Simpson is deciding whether they want "one expansive internal cloud or multiple discrete clouds dedicated to, say, Web serving or financial systems." (He's leaning towards the latter.)
• GE wants to "avoid building special GE code," and they "don't have the energy to figure out how to do this right all by ourselves." Simpson thinks they can "help cloud vendors get there faster."
• GE's IT organization currently charges business units a flat rate per physical server that covers power, cooling, and hardware. A private cloud would let them charge for IT resources based on consumption or per user.
• GE is evaluating cloud-based applications like WebEx, Zoho, Google Apps, and PostPath.
• GE has used Amazon's Web Services for "cloud bursting"
• CTO Simpson expects Cloud computing to reduce both power consumption and total employees managing the data centers
  

My take: if a company like GE takes cloud computing seriously, shouldn't you?

The way ahead tells us that the lines between the business and the consumer experience are blurring.  We saw that, for example, in the Bechtel case study we blogged about. 

Bruce MacVarish tells us a good portion of this story, and his diagram is a good indicator of where we're headed. 

Add to this the voice of our friend Tim O'Reilly, who's been telling anyone who will listen that "it's time to start thinking of the phone as a first class device for accessing web services, not as a way of repurposing content or applications originally designed to be accessed on a keyboard and big screen." 

I can't agree with him more. 

So where are we today? 

What does the cloud user experience look like at this point in our journey?

The basic promise of cloud computing: instant access to your data anytime, anywhere, on any device, is already a reality:

Here are some of the must-haves as we see it for the end-user experience, be it business or consumer:

• Upload multiple files of any type: (documents, presentations, PDFs, Website images, videos, music and more) Files of any type and size can be uploaded, stored and instantly shared.
• Real-time access to stored files, with trusted service providers hosting your data.
• Collaborate and share files with anyone--inside and outside an organization. Users must be able to securely share files with vendors, customers, family and friends, all without having to have an account or moving files from their stored location. Control of  customer data is where it belongs - with the customer.
• Hierarchical and tag-based file organization: hierarchical and tag-based file system organization and storing with library file views for easy search and retrieval.
• Permission Management: Role-based permission management and share expiration lets you manage your private and public shared data.
• Notifications: Account holders and non-account holders get notifications when files/folders are shared or when they have been assigned to a collaborative project. Notifications of shares to non-account holders leverage a viral marketing element to invite recipients to get involved.
• Publish-able URLs: users must have the capability to generate public URL links to share and publish files on forums, blogs and websites (of course, they must also be able to share files securely, in private).
  
• End-to-end security: data must be secure both in transit and at rest.
  
• Data security and control: use 256 bit AES encryption for each stored file with SSL encryption for files in transit. That takes care of the basic security fears for business users and consumers. Why should one be less secure than the other?
• Online restoration of deleted files: Users should be able to delete and restore files right from the web.
  

The good news? We can do that today.

And what's next? 

Here's what O'Reilly says about the future of mobile:

a sensor-rich device with applications that use those sensors both to feed and interact with cloud services. The location sensor knows you're here so you don't need to tell the map server where to start; the microphone knows the sound of your voice, so it unlocks your private data in the cloud; the camera images an object or a person, sends it to a remote application that recognizes it, and retrieves relevant data. All of these things already exist in scattered applications, but eventually, they will be the new normal.

Get ready. The cloud user experience promises to change the way we live and work.

Side note: We'll also be discussing the semantic Web and the requirement for tagging capabilities in cloud storage, but let's save that discussion for a future post dedicated to just that topic.

In his introductory post, Steve mentioned that when we were at VeriCenter we became aware of a new sort of challenge facing our industry - from outside entrants like Amazon.com who were beginning to grow their cloud storage and cloud computing businesses, leveraging internal web-based infrastructure. We knew then that cloud storage, as a significant part of an overall cloud computing strategy, was going to change our business and yes, our entire industry. Now we see new entrants like Google and Microsoft, as well as veterans like IBM and Sun, EMC, and Cisco are all delivering cloud-based offerings that are redefining the hosting and SaaS markets.

So here’s the question we’re asking: how will cloud computing disrupt your business model?

To answer this question, we start by defining a basic yet useful framework for understanding business models; this framework applies to businesses of all types and is adapted from Wikipedia:

Let’s look at the nine components of the business model framework, and describe each one briefly.

Infrastructure

• Core Capabilities: The capabilities and competencies necessary to execute a company’s business model. These include your key people, processes and technologies.
  
• Partner Ecosystem: The business alliances which complement your capabilities.
  
• Key Processes: The key activities (sometimes proprietary) which create the product or service you offer.
  

Offering

• Value Proposition: The compelling benefits your customers receive from buying your products and services.
  

Customers

• Customer Segments: The target audience for your products and services.
• Distribution Channel: The means by which a company delivers products and services to customers. This includes the company’s marketing and distribution strategy, and may involve a chain of intermediaries, each passing the product down the chain to the next organization, before it finally reaches the consumer or end-user.
• Customer Relationship: The links a company establishes between itself and its different customer segments.
  

Finances

• Cost Structure: The nature of the expenditures required to run your business, particularly considering whether costs are fixed or variable in nature, and whether they are capital of operating expenses.
  
• Revenue: The way a company receives payment from customers.
  
• Profit: The degree to which your revenue exceeds your costs, of course!