ENKI Blog (RSS 1.0)

Cloud Superman or Business Superstar?

2010-08-29T21:46:08+00:00

As we head into VMWorld week, the focus of my attention is going to be on all the cool new cloud technology that will shown off there, and what our competition has come up with. However, I've been talking to a number of prospects for our managed cloud computing services in the last week, and it struck me that none of them asked me anything about all the wonderful technology ENKI deploys into its PrimaCloud service. How can this be: the world is drowning in hype about cloud computing, and our next slate of customer partners don't care about cloud at all? (Of course, they DO care about what the technology can give them!)

At the same time, I was interviewing a potential VP of Bus Dev for ENKI who asked me to characterize our customers. This made me realize that the central question for our prospects is:

Do you want to be a Cloud Superman or a Business Superstar?

Our customers want to be business superstars. Most of them view the necessary evil of operations as something they haven't been able to get away from until they found ENKI, and that if they had their choice, they'd be focused on imagining the best way to grow their business 10x in the next year, not what the infrastructure needs to be to accomplish that, or how much knowledge of cloud computing they have amassed in the process. They know that their total cost of operations includes the consequences of all the mistakes one can make in managing software deployments in cloud or colo infrastructure. Often, they are experienced executives who have already made those mistakes, much like we did at startups and enterprises many years ago, and are not eager to repeat them.

We have a new customer, Krome Photos, who has an amazing outsourced photo-retouching/editing service that you can use to get what you previously needed a professional photographer to give you - but with your own pictures, and at a fraction of the cost. He gave us his entire operations responsibility, and we've expanded his infrastructure to keep up with his ever-accelerating growth. Now, he's preparing to charge for his beta product - almost unheard of in the Web2.0 world. I think of him as a Business Superstar!

To be sure, we get a lot of contacts from people who are worried about a 5-cent difference in CPU-per-hour cost, or whether we use Sun or NetApp storage systems, or whether they might lose some control over their production environment if they can't physically touch the servers or log in to the IPMI interface. Often those people don't become our customers or are impossible for us to provide our exceptional services to, since they aren't asking us to help them with what we do best, which is to take responsibility for their production IT deployments.

For example, we have one customer who thought his business was "too large" for a smaller cloud company like ENKI. But he wanted to try us out anyway to compare us to Amazon. So, he put his database in our cloud and his business logic layer into Amazon. The result is that if there is network congestion or Amazon suffers one of its regular "5pm doldrums", his application goes down. And he points the finger at ENKI. Unfortunately, as much as we'd like to, there is little we can do to save him from this situation because it's an architectural nightmare. He gets to be a Cloud Superman, but his business suffers. And he's not our biggest customer, but our smallest!

So, are you a cloud superman, or a business superstar? Drop me a line with our comment form below!

If you happen to be at VMWorld, come visit us a breakfast seminar we're putting on with one of our vendor-partners, Voltaire, which makes outstanding hardware as well as truly understanding cloud infrastructure. We'll be discussing "Designing Managed Clouds for Growth" on Wednesday September 1 at the Marriot next door to VMWare. Register at www.voltaire.com/vmworld-seminar

ENKI cloud IT services enables marketing seminar

2010-08-10T20:39:25+00:00

For a while now, I've been extolling the virtues of combining cloud computing with on-demand IT operations services such as system administration, software deployment, and incident response because it allows organizations to take on IT projects that otherwise would be beyond their skills or budgets. While we have many customers here at ENKI who simply purchase computer time from us in the cloud model, a growing percentage are combining it with our operations services (PrimaCare) to meet interesting challenges. Today I want to relate an amazing success story of one of our customers, Sitebuilder Corporation, who leveraged our fully-managed cloud IT service. First, a quick summary of what happened, and then the story:

To summarize:

SiteBuilder deployed a 350-seat website development network in a hotel with less than a day's work
PrimaCloud scalable cloud computing was used to deliver the development environment without any foreknowledge of required resources
Resources were measured using our monitoring system and adjusted on-the-fly to measure and match demand
PrimaCare outsourced IT services found performance problems and fixed them on-the-fly
Event was a success!
On completion of event, infrastructure is disposed of and re-created for next event so SiteBuilder pays only for what they use
Websites developed during the event can be migrated seamlessly to permanent hosting in the cloud
Customer accomplished the event with no permanent IT staff, no IT knowledge, and only a few days' preparation

The Story:

We've had a web design company, SiteBuilder Corporation, as a customer for a couple years now. They are so exclusive that they barely have a website, just getting business from referrals. Their primary focus is assisting their customers in building mass-marketing websites. Once those sites are built, they host them on shared infrastructure provided by ENKI, with operations services coverage so that their business doesn't need a 24x7 IT staff to handle the "distraction" of hosting. In fact, aside from a single programmer, our customer has no IT staff at all.

Recently they decided to write software that automated their web-design process, allowing anyone to easily build their own marketing website and start selling on the internet. They worked with a partner to teach people how to use the web-site-building software in the context of a 5-day internet marketing seminar taught at a hotel by a renowned marketing guru. After the class, students could move their website to cloud-based hosting and start their internet businesses. The class sold out quickly and they called us and asked us to provide a solution to hosting their interactive software for building 350 students' websites as they built them during the seminar. Their biggest concern was that 350 people all pressing "submit" at once would either bring down the server or saturate the hotel's network. They feared that to solve these problems, they would need to bring an unknown number of servers onsite, and wire up a network to run the class - something they didn't know how to do. The bandwidth problem was addressed by choosing a hotel with twin 30 Mb/sec links to the internet - and hence to ENKI's infrastructure. However, none of us knew what the seminar would do to a server.

ENKI responded by allocating a large virtual server in our VMWare-based PrimaCloud product, which can scale the server on-the-fly to match demand. We knew that we could size the server up to the full capacity of the physical hardware, which was 24 cores and 96GB of RAM - if we needed to. We set up a 4-core, 16GB server for the customer, pre-installing the Wordpress framework on which their software operated, and testing it. We also worked with the customer to do some load testing, which looked good. However, we all knew that actual students would exercise the system in ways that the load testing could never achieve. In addition, this was the first major deployment of the web-site-creation software, so we had to be ready for anything. So, we applied our enterprise-grade server/application monitoring on the server which allowed us to see how busy it was, and what was going on inside the application software - in this case Apache and MySQL.

Yesterday, the seminar began, and our technicians watched the load on the server rise slowly and then peak when the students all submitted their "homework." Disaster! The load on the server climbed to 200 (the number of programs waiting to run) and all 16GB of RAM were used up. So our team upped the allocation on the server to 32GB and 8 cores, and still it wasn't enough - we got frustrated calls from the event saying it was too slow. So we upped the resources again, and the class proceeded, albeit more slowly than they had hoped.

Our operations services staff was looking at the server the entire time, and saw that there were tremendous numbers of open database connections with uncompleted queries causin lots of Apache threads to lie around and use up CPU and memory without getting anything done. So we sat down with our customers' developer to look at their code, and we also examined the database transactions to see why the queries were not completing. What we found was that both WordPress and the customer's code were generating queries against improperly indexed tables in the database. A few small changes to the database setup were made and today, the class is proceeding along snappily with just 4 cores and 10GB of RAM, with a server load of 0.5! Because we had been able to profile the customer's software in real time, we were able to fix performance problems and right-size the cloud deployment to save them money.

Our customer's partner is planning more seminars in different cities, and they will be paying only for compute and operations services during the time then need them.

When I realized how well this was working, I was stunned by how our outsourced cloud IT model had enabled our customers to accomplish something that had previously been impossible with their resources. In the past, they would have had to hire an IT organization to go to the venue to build a network, set up multiple-server infrastructure, and deploy the software. Moving the seminar from city to city would have been difficult and very expensive, involving a large staff of technically capable "roadies."

Today, we received a text from our customer, saying "Your help turned me from public enemy to rock star over night ... I thank you for your effort and expertise from the bottom of my soul!!!!"

A Beard and a Dog

2010-06-11T23:50:22+00:00

Today I was eating lunch at a sidewalk cafe with my dog at my feet, and a man walked up to me and asked me if I could help him. He looked sort of like a slightly disheveled yet surprisingly well-dressed homeless person, so I was curious what he wanted and asked him how I could help. He seemed confused and asked me what street he was on and where some other streets in the area were. I told him and asked him what he was looking for, and he said "my car - I lost track of it two days ago!" He then went on to say that he had Multiple Sclerosis and it made him forgetful, but that he could trust me with helping him since I looked trustworthy - because I had "a beard and a dog."

I was surprised and amused to hear him say this, but it made me think: what are the clues that we use to decide if we're going to trust someone? And more importantly for you and me, how will you decide if you trust us at ENKI to serve your business?

For me, I most often trust someone based on my intuition, and then secondary characteristics like their track record, their affiliations, their experience, knowing who trusts them, or simply knowing them in some way perhaps because we went to school together or worked together.

Since ENKI provides not just cloud computing but also operations services that effectively take the place of all or part of an IT department, it is critical that we be trustworthy, and that prospective customers trust us. Our business was founded on providing trust as the fundamental product, and our Enki Way defines how we go about it. But customers who come to us still need to know more than our intentions in order to trust us. Typical ways I see them trying to figure this out is asking us:

How long we've been in business providing cloud (4 years: longer than any other cloud computing company.)
Getting references by referrals (our #1 source of new customers is referrals!)
Checking the company principals - such as me - out through social networks, bios, associations, etc. (For example, we get quite a bit of business from Cornell Alumni who are entrepreneurs and businesspeople that I meet through my Cornell connections.) Our founders' experience as the CIO of NetSuite has definitely brought trust to us as having a team that knows about reliable computing.
Looking at the terms of the relationship that we offer, such as our 10x downtime guarantee, our rapid turnaround service SLAs, and our satisfaction guarantee - all of which show that we have skin in your game.
How many customers/employees/locations we have.
What technologies and vendors we use, presumably to gauge our judgment.
and more...

These are all passive methods, in the sense that people use them without interacting with us. No matter what answer we give, this makes ENKI look a lot like other cloud providers (aside from the technical and service differences of our offering.) In some ways, they are like looking to see if we have a beard and a dog! Personally speaking, there are plenty of guys with beards and dogs that I'd find difficult to trust.

Instead, what we have found is that our best tool to convince people to trust us is... us. The majority of our customers' executives have met our founders at some point, even if they're based in other countries. We've even made early and late trips to the airport to meet customers on their layovers. And failing a personal meeting, a phone call with our extraordinarily gifted sales team or with one of the founders can tell you a lot about us. Each of our founders and partners is always happy to talk to propects about ourselves and of course, your challenges in securing cost-effective enterprise-grade computing for your business.

So if you find what we offer interesting, I invite you to call us. I personally will be delighted to speak with you. Even if you aren't impressed by beards or dogs!

By the way, I never found out the particulars of why that man was looking for his car. It was a strange story - strange enough that I will keep looking for him every time I go downtown.

Why Cloud hasn't made "going down" a thing of the past

2010-05-05T17:20:21+00:00

Yesterday I read about how Foursquare had major downtime. They're a customer of Amazon AWS's cloud, yet they had downtime comparable to the conspicuous disaster in 2007 at 365 Main that took down non-cloud-hosted Craiglist, Yelp, Technorati and SixApart for up to day. What happened? Wasn't cloud computing supposed to solve this problem?

in this case, there was apparently a problem in the AWS datacenter that hosts Foursquare's services. However, as in the case in 2007, Amazon's downtime was far less than that of Foursquare. This brings up two myths of cloud computing: 1) Cloud computing never goes down; and 2) Your site will never go down if it's hosted on cloud computing. Both are terribly and dangerously false.

Hardware will always fail, suffer from misconfigurations, or otherwise be the victim of human failures, so from the user's point of view cloud computing will never reach 100% availability. In fact, when Amazon says that it's cloud is reliable, they speak of their entire cloud, not a particular end-customer's application hosting. The key is what your cloud computing provider does when the hardware fails. Amazon does very little, simply offering you the ability to restart your software on another machine. Others - such as us - will restart it for you. How quickly the restart happens, and how much demand on your software that it presents depends on the cloud you choose.

But that's only the beginning of true application reliability. As we saw with Foursquare yesterday and the 365 Main event in 2007, if your software isn't written and deployed in a way to be tolerant of failure, the time needed to bring it back up can be a major disaster for your business. This preparation for downtime is called DR, or disaster recovery, and people realized that the 365 Main event was highlighting that many companies haven't given it much thought. DR can start with as simple a preparation as writing your software to be resistant to database corruption caused by downtime or perhaps adding monitoring that shows whether it's working properly, or it can extend all the way to having a live "warm" standby site that can take over if your primary site fails. Cloud computing can make these options possible or affordable, but it does not guarantee them. So simply placing your site in a cloud doesn't guarantee uptime, even if it does put it in a datacenter that is professionally managed.

Every day, as i meet potential customers, read advertisements from other cloud companies, or catch up on cloud computing blogs, I've been making a mental list of the "myths of cloud computing" that I've been hearing from them. These myths are dangerous: they produce a mismatch of expectations between cloud customers and vendors that can injure everyone - especially the cloud user who expected that their cloud-hosted web site would produce far more professional results than it actually does.

A partial list of the myths I plan to explore are:

Cloud computing never goes down (I'll follow up on this article in more depth and make it part of my Cloud 101 class.)
Computing resources in the cloud are infinite
Cloud computing is nearly free
Cloud computing will make my software compliant with regulations and certifications
if you have any more you'd like me to discuss, I'd love to hear from you.

Like any myths, there is a kernel of partial truth to these assertions, which is the reason cloud computing is so attractive. But how much of these benefits you actually get depends both on you and your cloud vendor.

What is Cloud 2.0?

2010-04-12T22:18:19+00:00

ENKI believes the future of cloud is increased usability, availability, and performance that will enable true enterprise computing. The key is combining application-based operations services that companies have had to provide for themselves until now, with on-demand computing. We call this Cloud 2.0.

When ENKI first entered the pay-as-you-go, on-demand, managed computing services market 4 years ago, the term "Cloud Computing" wasn't even in common use. Since then, Cloud Computing has come to be commonly accepted to denote on-demand computing delivered over the Internet. Businesses using applications in the cloud (SaaS, or Software-as-a-Service) have learned to expect that they will not need to add any staff to manage those applications. However, companies that want to use cloud computing to replace their in-house or colocated datacenters for production computing (IaaS, or Infrastructure-as-a-Service) have not been seeing the dramatic cost redutions promised by cloud vendors who only replace physical hardware with remote virtual hardware. Instead, they have simply replaced staff who manage physical servers with staff who manage cloud servers.

ENKI has four years of experience delivering cloud computing that begins with outsourced virtualization, but adds operations services including system administration, security services, incident response and software release management. By offering SLA-focused management for their cloud deployments, ENKI enables our customers to enjoy improved service levels while reducing their management costs using the same on-demand model as for their computing.

We call the combination of cloud computing with service-level focused management, Cloud 2.0, as it fulfills the original promise of cloud computing to not only deliver improved services but also overall cost reductions.

InfiniBand and You

2010-03-29T23:59:25+00:00

Written by: Louis Corso

As the demand for faster interconnect continues to push on the limitations of Ethernet and Fibre Channel technology, InfiniBand's time to shine is coming soon. InfiniBand provides some unique advantages of legacy technologies. Though InfiniBand is not on the radar of many data center managers and IT directors today, it stands to gain ground as IB switch manufacturers announce throughput numbers this year which are beyond anything seen previously.

InfiniBand has been on the radar here at ENKI for a few years now. Our CEO Dave Durkee saw early on that the real bottleneck in a virtualized environment was the connection between the storage and compute layers as well as collaborating virtual machines in virtual private datacenters. He knew that the 10 Gb/s promised by coming technologies in Ethernet and Fibre Channel, while adequate for physical environments, would not cut it once a hypervisor was layered on top of the hardware since each virtual machine could nearly saturate such a pipe. For a true high performance, enterprise-grade cloud experience, Dave saw InfiniBand as the clear winner for several reasons. With a number of recent blog posts and press releases touting the benefits of IB, we have seen a growing number of prospects lining up at our door due to Dave’s foresight. If you don’t know about InfiniBand, now is the time to learn, and ENKI is here to help.

While researching for this post, the first place I went was to the website of the InfiniBand Trade Association. I encourage you to do the same, but I’ve reprinted some of the highlights taken directly from their site here as well:

- InfiniBand is the only publicly available solution to support 40Gb/s host connectivity and 120Gb/s switch to switch links.

- InfiniBand offers ultra-low latency, with delays measured as low as 1µs (.000001 seconds)

- An InfiniBand link can carry networking, clustering, and storage data over one connection for simplified management

- InfiniBand allows for fully redundancy with automatic failover to meet the demands of mission-critical applications

- InfiniBand has a detailed roadmap in place for increasing connection speeds over the next 2 years. Within the next 3 years, bandwidth approaching 1000 Gb/s is expected

Even with the increased performance it provides over legacy technologies, InfiniBand is still little utilized outside of the HPC community, where it is the chosen method of interconnect in five out of the top ten clusters, as of their November 2009 report. IB remains mostly an unknown throughout much of the enterprise IT industry. According to a recent survey by IB switch manufacturer Voltaire, 54% of IT industry respondents intend to rely solely on Ethernet for their future data center expansions. All this may be about to change, however, with the release of a new report by Voltaire claiming that their Grid Director 4036E switch, featuring 34 40 gb/s IB ports, delivers a full 2.72 tb/s! With transfer rates busting through the ceiling, I believe that the data center community will have a hard time continuing to push InfiniBand into the margins.

Going forward, I predict that we will see more manufacturers announcing I/O results that were previously unimaginable thanks to InfiniBand. ENKI intends to ride this wave as well, and I expect that you will see some interesting test results in the near future supporting Dave’s decision to implement IB in our data centers. New hardware such as the 4036E, which is currently powering data through ENKI’s PrimaCloud IaaS platform, blows current and future upcoming generations of Ethernet technology out of the water, something that will be hard to ignore as developers continue to challenge the data center to do more and more.

Cloud Computing: the New Paradigm of Enterprise IT (part 1 of 3)

2010-02-25T00:44:40+00:00

Written by: Dave Durkee (with Louis Corso)

In the timesharing model of mainframe computing, each user was granted a fixed slice of the resource pie (compute and storage) with which to perform tasks. The system kept a record of the actual resource usage of each terminal, which was then printed out at the end of each session. The report detailed the following: CPU seconds, storage I/O ops, and total number of storage blocks used. Timesharing proved to be a very effective way to grant computer access to multiple users, at a time when computing was a very expensive endeavor.

When I was in college, in the late 70’s and early 80’s, computer timesharing was the dominant method my fellow students and I would use to access the campus mainframe, which was locked away somewhere in a university data center. We would go to one of the 30 or so terminals in the computer lab, each of which was connected back to the mainframe. From our individual terminals, we were each granted a small time-slice of the mainframe’s CPUs to write programs, run them, and play games.

Gradually, businesses began to recognize the value that timesharing provided, by allowing many users to share one computer. A new type of company was born, the Service Bureau, whose sole purpose was to provide computer timesharing services to other companies. This represented the first major adoption of computing in the corporate world. Over time, Moore’s Law began to kick in, reducing the price and complexity of computers to a point where businesses could afford to invest in their own hardware. The need for Service Bureaus quickly began to diminish.

By the mid 1980s, just as quickly as they had arrived, most of the Service Bureaus were gone or had moved on to other ventures as companies absorbed the functions they once provided. This transition, from externally hosted third party computing to internal data centers, represented a major shift in consciousness by the entire business world. Suddenly, companies began to realize that this “thing” called a computer was really good at doing many of the tasks businesses needed to complete on a daily basis. One example of this, payroll processing, was turned into a very efficient computer-automated process by companies such as ADP. As time passed, businesses came to depend on computers for an ever increasing share of their day-to-day operations.

The story of the last 30 years of corporate computing has played out as a series of paradigm shifts like the one described above. During this gradual evolution process, companies have retained the lessons learned from previous technologies and paradigms, such as ERP and timesharing, while continuing to move forward to meet the needs of an ever more challenging business climate. Learnings of the past are now taken for granted by a business community with its sights set on the here and now.

Each of these shifts was the result of a new business need, and a corresponding technology which was best able to meet that need. Like those times, now is a time of great challenges for business, with needs that a different than those of the past. Through this series of posts, I intend to outline the new problems faced by business users in the current IT landscape, the ways that cloud computing and outsourced IT addresses these problems, and the other residual benefits cloud users stand to gain. As you read, keep in mind the challenges faced by your organization, and how they can potentially be solved in the cloud.

Price-focused v. Value-based IaaS Vendors

2010-01-19T00:32:47+00:00

Written by: Louis Corso

Amazon’s new auction based pricing strategy brings to light some philosophical differences between the price-focused and value-based cloud vendors. While they are ideal for non-mission critical tasks such as software development, price-focused vendors are less prepared to meet the needs of a growing demand for enterprise grade performance and reliability.

Moving your Development and Test to the Cloud

2009-10-09T08:00:00+00:00

I got an email yesterday from some cloud vendors offering a seminar on "Why you should move dev and test to the cloud." The two vendors specifically suggested moving them to the cloud even if you weren't deployed to the cloud. For our cloud customers, I have been advocating this for quite a while, but I would never suggest moving test, and to a lesser degree, dev to the cloud if your deployment platform wasn't the cloud. My disagreement is based on the importance that I put on keeping your live site or other service up and running.

The real recommendation I have, based on our experience supporting our customers, is that your dev and test environments should be as much like production as possible (with any obvious required differences such as different levels of access control.) So, if you're deployed to cloud, you should test in an identical cloud environment. If in colo, you should test in an identical colo environment. That way, any environment-dependent problems will surface during test, instead of during deployment. And we find that more than 50% of the failed deployments that our customers have are the result of having a difference in their environments, whether it's a different PHP config file, different memory size, or different set of libraries on the systems. This applies to development as well.

Yes, the cloud offers some great advantages for "disposable infrastructure," letting you set up and destroy test or dev environments at your leisure. But if their use leads to a bad deployment and downtime, any advantages of cloud dev or test are lost. The nice thing about our AppLogic-based cloud or PrimaCloud is that you can copy an entire virtual datacenter from your production environment and make it your test or dev environment very easily.

All of this ease of creating and destroying infrastructure doesn't eliminate the need for good process. Do you have configuration management and version control tools in place? Do you have a process with approvals for releasing from dev to test? How about from test to your production environment. How do you decide how much testing is enough, and of what type? What process is in place if you discover critical bugs after release? Can you roll back your production environment easily? The cloud facilitates answering these questions, but as you can surmise, it doesn't answer them!

Larry Ellison Doesn't Hate Cloud Computing

2009-10-01T19:34:25+00:00

Hello! My name is Louis Corso, a new sales rep who joined ENKI to bother all of you. Hopefully you and I will become great friends in the way that this social media stuff intends us to, but in the mean time, here are a few interesting facts to keep in mind.

- I'm a 24 year old Silicon Valley resident... depending upon who you ask, this is either a good thing or a bad thing... the way I see it is this: I grew up immersed in IT because of my age, family ties, and home town (Cupertino, CA). This is never bad.

- I love Apple products. I've sold them, I've used them, I've forced other people to use them, including COO Eric Novikoff, who will be buying his Macbook Pro soon enough.

- I am a sports nut. If I wasn't talking to you about IT right now, I would be talking to you about the San Jose Sharks, who play tonight, and whose jersey is draped over my chair.

Now that we have that out of the way, let's get down to something more interesting, namely this:

TechPulse360 posted this video on YouTube earlier this week. In it, Larry Ellison presents some insights about cloud computing, along with some pretty funny nonsequetous rants. I figured I would talk about the highlights. :)

0:10 My objection to cloud computing is... Cloud computing is not only the future of computing, it is the present, and the entire past of computing is all cloud.

Thank you Larry! This statement really gets to the core of what this whole new way of thinking about computing really is... just computing. The next quote will tell you what I mean by that.

1:25 It's not water vapor. All it is is a computer attached to a network... It's databases, and operating systems, and memory, and microprocessors, and the internet.

I got a laugh out of this because we have been having these kinds of conversations behind the scenes at ENKI for some time now. So many times, we have talked to customers who are stuck in the rut of thinking that cloud computing represents something foreign. In reality, our data center has the same stuff in it that everyone else's does. Once people realize this, we can stop being scared and start talking about the real problems cloud computing can address.

2:55 Are we dead by cloud computing? Is Intel dead? There'd be no microprocessors in the cloud? Is Samsung dead? There'd be no memory in the cloud? Is Cisco dead because there's no networking in the cloud? Are we dead because there's no databases in the cloud, no applications in the cloud, no middleware in the cloud? The answer is no. All a cloud is is computers in a network.

As I sat down to write this article, I didn't realize that there was a core theme running through each of these quotes. I have more, but I don't really think I need to share them for you to see what Larry Ellison (and I) are trying to get at: the cloud isn't nearly as scary, or foreign, or new as many people think.

3:25 In terms of business model, you could say, ‘Oh, it's rental.'

I read a really great quote on TechCrunchIt's forum about this video, credited to Sean Mahoney: "It would be like asking a farmer if he is worried about the new trend of ‘restaurants.' People still eat, even if they consume in a different way." What we are offering our customers is a different way, I think a better way, to ‘consume' the same thing they've always needed.