Sunday, August 21, 2011

Why cloud computing hype isn't bad for IT after all

A week or two back, ReadWriteWeb ran and published the results of a reader poll of the “most over-hyped cloud technologies.” Amusingly, the results (aside from a NoSQL mention) read like the basic NIST definition of the key components of cloud computing. Software as a Service, private clouds, Infrastructure as a Service, and Platform as a Service all made the top 5.

Wow, I thought. That barely scratches the surface. Plenty more cloud computing terms were enjoying their moment of irrational exuberance, but were being left out in the cold by this particular survey. A few Twitter conversations unearthed some very deserving nominations. Not to be forgotten:


Hybrid clouds. Apparently hybrid clouds didn’t have quite enough hype-y-ness to make the list. Weird, considering that term tends to be the punch line to nearly every cloud strategy and direction conversation that I hear. Better luck at next year’s awards ceremony, I guess.


Cloud bursting (nominated by @reillyusa and @AaronMotsinger). Some folks have been arguing back and forth about whether it is really a legitimate (or even possible) use case. @pdowning1077 noted he much preferred the term “capacity on demand,” but that doesn’t help settle the argument.


Cloud brokers. Forrester has been posting some interesting research for its subscribers on this new role (also defined by NIST in the July 5 version of its standards roadmap, if you want a standards org to weigh in for legitimacy). I’d say this conversation is still very early. The hype wagon train for this term has just set off down the road.


But probably the most impactful comment was another by @pdowning1077. “How about just the term ‘cloud’ [in general]?” he asked. How could they forget to include the mother of all hype-worthy terms in their polling?


So much hype that “cloud computing” becomes meaningless?


The same week of all this discussion, David Linthicum reported that cloud computing (the term) had now become essentially “meaningless.” That comment came on the heels of Gartner’s annual publication of their hype cycles. A quick scan notes that cloud computing is still close to where it was last year, just nosing over the (hype-laden) peak of inflated expectations. Private cloud computing is rapidly moving to join it, perched perilously over the trough of disillusionment, ready to take the leap.


OK, no one would argue with the extreme levels of marketing attention from everyone from start-ups to 30-year-old software companies (who, us?) to service providers. But just because a bunch of marketing people are in a frenzy doesn’t mean we should write off the trend they are talking about as a bunch of meaningless fluff.


The hype has caused IT to pay attention to cloud computing


In fact, if I’m reading the market right, I’d say that there are actually some really good things that have come out of the hype around cloud (and continue to do so).


We suddenly had something to call this good idea. There were a bunch of technologies and entrepreneurs out there struggling for several years to put a palatable name to what they were working on. Some started off calling this grid computing, some utility computing, and others more obscure terms than those. But, the early hype around cloud computing a few years back gave a name to the idea. We pulled several of these companies into CA (Cassatt and 3Tera, to name two), but many others were struggling with this same issue. One of my early posts on this blog was about how the term private cloud may not have been precise or perfect, but it enabled us to have the right conversation. I think the same thing goes for the overall cloud computing concept.

It created a way to catch the attention and imagination of enterprise IT. By talking about a Big Vision of IT infrastructure that matched compute supply with compute demands at any given time (and matched costs accordingly), ears perked up. It was the next logical topic to discuss with the IT guys who were fresh from thinking about how virtualization could free them up from particular pieces of hardware. In a world in which IT is fighting for every budget dollar, mostly just to keep treading water, an idea about how to get off this downhill hamster wheel is at least appealing to consider. That’s step one. (Ken Oestreich, by the way, has a great blog from a few months back on the brief history of the vision of cloud computing.)

The hype extended the discussion past the technologists to the business people. All the hubbub over cloud computing got the business users excited at a time when the economy was giving them little to be excited about. “So, you mean I might have a way to turn some of these business ideas into reality, despite the drubbing that the sour economy has given us and the measly budget that my IT partners say we have at our disposal?” This has been important – the business guys are the ones, in the end, pushing when IT starts to get nervous and pulls back from the visionary edge that cloud puts them on.


The hype has pressured big vendors into some self-reflection that will be beneficial for their customers. Many of the larger vendors jumped on the cloud bandwagon through new offerings, blatant rebranding of old offerings (shame on you), acquisitions, and the like. To make any of these moves, vendors have had to take stock and rethink what they can and should be providing given what their customers want. In some cases (like here at CA with Nimsoft), it causes the vendors to broaden the set of customers they are actually serving.


The intense amount of discussion has started an intense amount of scrutiny, revealing how useful cloud can actually be. One thing that happens when the hype levels reach fever pitch is that people start pushing back. The recent demand for real-world examples and exasperation over cloud outages has been the natural backlash from being force-fed lots and lots of best-case scenarios, rainbows, and unicorns. Journalists and analysts have often helped push for these kind of reality checks, though they also tend to pile on as technologies or ideas drop into the “trough of disillusionment” that Gartner is so fond of describing. Enterprise IT, business users, and the vendors themselves all eventually do a fair bit of policing, sometimes too late for their own good, but we seem to be headed in this (positive) direction right now.


So while a lot of the hype can seem like so much wasted energy from all parties, when the trend or shift being hyped actually has merit, something useful comes out the other end. Now, would

most of us (advertising agencies and ad reps excluded) prefer some way to skip the aggravation of this process and jump right to the end? I’d bet so. However, consider this all a bit of a trial by fire. The only way for something to be proven strong enough to pass through the fire is, well, to actually do it.


So, hold your nose and smile. Hype is good – with a few important caveats. Be critical. Be well-armed with the right questions to ask in order to discern the valuable from the merely fancifully over-marketed. Be ready to see the value in approaching something a new way, even if it’s something you’ve done the same way for decades. Be pragmatic enough to know it won’t happen overnight or with the wave of a magic wand.

If it makes you feel any better, cloud computing isn’t the only term getting the Gartner Hype Curve treatment this year. Added to the list, according to this ReadWriteWeb article, were big data, gamification, Internet of Things, and consumerization. Misery loves company, I guess.


And, in the meantime, it may be time to come up with your own term to start campaigning for next year’s Cloud Hype Awards. I think the hype is here to stay for a while longer.

Monday, August 15, 2011

Wallace and Bird Hosting soar: how they save on hardware while delivering more complex apps

Many of the cloud service providers that we are working with here at CA Technologies are small, but aggressive. These are the guys that know their business really, really well and are the ones who are targeting their niches pretty successfully. The ability to focus and be nimble are some of the key indicators of success in the service provider space.

We’ve been calling folks like these “Cloud Accelerators” – given how instrumental these service providers are being in the adoption of cloud computing -- and we have been profiling some of the more interesting ones.

Bird Hosting strikes me as a great example of one of these “accelerators.” They may not be on your radar screen yet, but Bird Hosting is a cloud service provider that’s nimble and all about the personal touch. They not only deliver cloud hosting and other key services to customers, you can find them answering questions on WebHostingTalk.com. They post reviews talking through the pros and cons of the newest rev of their cloud platform. They even run a hardware review site. (After all, there are benefits for a service provider to know this stuff.)

They do all this with 10 employees, 3 locations…and they have 3,600 customers.

Bird Hosting’s CEO, Michael Wallace, answered a few questions about their operations, what’s key for a service provider’s cloud platform, and even what’s up with their name. Read on:

Jay Fry, Data Center Dialog: You’re a relatively small service provider. How many data centers do you need to effectively serve a nationwide customer base? What else is important?

Michael Wallace, Bird Hosting: We started in Seattle. When you have a data center in Seattle and customers in New York, the content isn’t delivered as quickly as you’d like – content delivery speed is important. We saw a need for something more central in the U.S., and so we went to Dallas. We now have 3 data centers (Seattle, Dallas, Washington, D.C.) and can cover the whole United States effectively.

In fact, we have customers across the world. We have customers that range from small shared-hosting accounts to large corporations. For example, we support a travel organization in New York, a food chain also in New York, as well as Costco. We host and produce videos that have 360-degree views of sheds and playgrounds that Costco sells.

DCD: How did you get started in the service provider business?

Michael Wallace, Bird Hosting: We originally started in 2005. My father was one of the pioneers of the Internet; he started an ISP in 1991. I started out doing web hosting with some big corporate website hosters. I found that they were subpar, both from how they presented themselves to the services they offered. So I decided to do my own. I had the knowledge to do it. And I had just returned from a lengthy tour of duty with the U.S. Air Force in Iraq and was ready to take a bold move and launch my own business.

DCD: A recent article in SearchITChannel by Carl Brooks talked about the fact that the market is starting to realize that enterprises who are building private clouds and service providers that are selling public cloud services have some very different needs. Can you talk a bit about what you think is important to have in a cloud platform for service providers like yourself?

Michael Wallace, Bird Hosting: Two things that are very important are cost savings and flexibility. With CA AppLogic [what Bird Hosting uses], I’m not limited. It’s a platform that allows me to offer a full range of services. Before I had to have separate machines for each individual task. We had to have a box for Windows, a box for Linux, and when we added customers, we had to add more dedicated boxes. With AppLogic, you don’t have to be limited to a certain OS. You have a cluster of servers and it allows us to run all of our applications on top; it’s neutral. So that helps us save money – we are able to standardize hardware and we can even save power.

DCD: Speaking of saving power, I’ve heard you talk about how much you’ve reduced your environmental impact. Can you quantify what you’ve been able to do?

Michael Wallace, Bird Hosting:
When we first started our business, we sold a lot of dedicated servers. Each would have a power draw even when idle. CA AppLogic allowed us to take those boxes and cram them into a virtual environment. Customers on dedicated servers were able to run as virtual dedicated servers instead. Customers were paying for a given amount of resources; we would give them the resources and they would run their environment. We cut back from 6 full cabinets to just one cabinet in one data center.

We were able to take 200 boxes and put that into 10 boxes. In our Dallas facility, for each of our cabinets we have two power drops, which costs $960 per cabinet – just for power. We were able to cut back 10 power drops, saving us about $10,000 per month.

DCD: Service providers are all looking for ways to define a niche that they can own, build up new revenue opportunities, and grow their margins. What has your cloud platform decision and approach meant competitively?

Michael Wallace, Bird Hosting: It allowed us to broaden our market. We could have stayed where we were and offered shared hosting and dedicated servers. But using AppLogic has allowed us to venture into more complex application hosting. It allows us to create more complex environments with ease. Those environments usually would have taken us a long time to set up if we were going to do it in a physical environment—running cables from an end gateway or a firewall to a hardware load balancer to machines. That’s what AppLogic does in software.

We can support more complex applications. We wouldn’t be able to offer as many value-added services without AppLogic. We can do so quickly and efficiently, scaling things up or down as much as they need. And then we can charge per resource.

DCD: Do you see IaaS providers like Amazon as competitors?

Michael Wallace, Bird Hosting: I don’t consider Amazon a direct competitor – they can’t do what we can do. We have a running joke – customers come in and say “our uptime is 70% with Amazon Web Services.” We can give them close to 100%.

DCD: You also do a lot of stress testing with your hardware. What has that meant for customers? What have you found?

Michael Wallace, Bird Hosting: CA AppLogic is designed to take a bunch of small systems and merge them into a cloud. With more cores and more memory on each machine, you can get more out of each machine without adding more hardware footprint. We decided to start a little testing to find out what are our limits, what can we do? We tested different types of hard drives and configurations, different controllers, to see what worked best.

Testing the efficiency of servers brought me into a whole new area. I now run a hardware review site on the side. The stress-test work helps us figure out what machines can handle without impacting performance.

[Note: you can contact Michael directly to talk about any of his results if you’re interested in more details.]

DCD: Do your customers know anything about cloud computing? Does it matter to them?

Michael Wallace, Bird Hosting: There has been a lot of confusion and debate with the cloud. When I started out, we heard concerns about sharing data. We see less concern about that now. They do see the power of cloud computing and interact with it if they are using our virtual private data center offering. We dedicate a given group of machines to a customer and they get access to the CA AppLogic portal so they can build applications the way they see fit. They use CA AppLogic to virtually build their infrastructure.



Thanks for the time, Michael.

Oh, and in case you were wondering about the name, you’ll be happy to hear that I (of course) didn’t let that go. Here’s what Michael said about the name: “When we started looking for names, I was also looking for a mascot. That’s where Bird Hosting came from.” Did he try other more, er, ferocious names? “HostGator was taken,” said Michael. “And Dog Hosting was too gangster for me.”

The resulting mascot has an uncanny resemblance to Twitter’s feathered friend, if I do say so myself. Or the stars of a certain iPad game I’ve gotten myself hooked on.

But no Angry Birds here. Just a set of happy customers benefiting from Bird Hosting’s personal focus on delivering what experience dictates is required for service providers to soar: performance, reliability, and value.

You can read the Cloud Accelerator profile of Michael and Bird Hosting on the CA Technologies site. The site also features other luminaries who are setting the pace for cloud computing.

Wednesday, August 3, 2011

Boy, my new iPad and I are demanding a lot from IT -- and we're not alone

I caved and joined the revolution this weekend. I bought an iPad.

And while it was very fun to do all the things that a newly minted Apple fan boy does (like downloading the app that turns the iPad into ones like they had on Star Trek: The Next Generation), that was just the beginning. I had yet to try to torment my internal IT department with my demands.

First and foremost: I wanted to use my iPad as part of my normal work day. I'm certainly not the first to want this. The appearance of consumer-purchased devices that employees would like to have (must be!) supported by internal IT is getting an amazing amount of attention. Though not always from IT departments, if they can help it. In addition, it’s not just 1 or 2 folks who want to start using tablets, smartphones, and the like. It’s everyone.

What does “not supported” mean for my iPad?

So, first thing Monday, I tried my luck linking into our IT systems. It started off innocently enough: I easily connected to the in-office wireless network. The first real test was going to be whether I could get my corporate email and calendar.

Dilbert.com

IT had obviously been through this before; there is a document in place on our help system that explains how to do everything. Unfortunately, it starts like this: "Please check if this iPad was purchased for business purposes or if it was a personal purchase. Note: personal machines are not supported."

Hmmm. That sounded ominous. But, despite being “not supported,” it was really simple to enable email and calendar access. I had to add some security precautions, as you might expect, but it worked. My fingers are crossed that it continues to work, given the help I’m not going to get. And, of course, there are a multitude of our enterprise apps I’m not getting access to.

But I’m satisfied. For now. But not everyone is. And corporations certainly shouldn’t be.

Cloud computing, intriguing mobile devices (and what you can do with them) are ganging up on IT

My process of tormenting IT with my iPad started Monday, but it’s guaranteed to last for a long time. And, as I said, the key issue is that I’m not alone.

People – and, yes, it’s about the people and what they (er, I) want to do – have devices that they love that give them easy, constant access. That should be good. There’s a blurring of the boundary between business and personal that businesses stand to gain from.

Cloud-based resources give organizations a fighting chance to scale sufficiently to keep up with the volume driven by these more-and-more-ubiquitous tablets and smartphones. But management and security are often thought of way too late.

In a piece posted at Forbes.com, Dan Woods, CTO and editor of CITO Research, noted that “the IT monopoly has ended but the need to ensure security, reliability, compliance, and integration has not. Most consumerization efforts are long on choice and short on ways to actually address that fact that IT’s responsibilities to manage the issues have not disappeared.”

Shirking management and security – or leaving it as an afterthought – will not cut it this time around, especially since users don’t think twice about going around the official IT channels, something that those official IT channels really can’t afford to have happen if they are going to get their jobs done.

The train is moving faster than you thought

In a study called “IT Consumers Transform the Enterprise: Are You Ready?” that IDC published a few weeks back (free copy without registration here; CA Technologies was a sponsor), they mention these needs – and the speed they need to be dealt with. “The train is moving faster than you thought. Adoption of public cloud, mobile, and social technologies in business operations has already reached high levels, often driven by ‘stealth IT.’”

IDC noted a “surprisingly high” (and concerning) level of personal and confidential information sharing. While the “consumerization of IT” introduces a bunch of new, innovative services and approaches into the enterprise, it also exposes the org to “business risk, compliance gaps, and security challenges if they are not managed.”

An InfoWorld article by Nancy Gohring noted another IDC study that found that even as more and more people are bringing their own tablets and smartphones to work, IT departments have been “slow to support them and may not even be aware of the trend.” Slow, I understand (given I just bought my first iPad a few days ago); not aware, however, is a recipe for big headaches ahead.

What are those ahead of the train doing to stay ahead?

Not everyone, however, is behind the curve. Part of the IDC survey I mentioned earlier highlighted the top characteristics of leaders in this area – as in, what behaviors are they showing. The leaders are more likely to be those using IaaS, PaaS, and Saas; those who are interacting with customers using their smart mobile devices; those who are concerned about data protection, back-up, and end-to-end user experience. “Businesses that are being proactive about consumer-driven IT are more likely to realize greater benefits from investments made to address the consumerization of IT,” said IDC’s Liam Lahey in a recent blog that summarized their survey findings.

In addition, in Woods’ Forbes article, he pointed out some questions that need asking, many at an application level: “Supporting mobile workers adds a new dimension to every application in a company. Which applications should be supported on mobile devices? How much of each application should be available? When does it make sense to craft custom mobile solutions? How can consumer apps become part of the picture? What is [the] ROI for mobility? How much should be invested[?] None of these questions have obvious answers.” Another post of his has some good suggested approaches for IT.

My CA Technologies colleague Andi Mann did a good job of netting this all out in another recent post: “While a minority of leading organizations already ‘get it’, there is still a massive latent opportunity to establish new game-changing technologies, drive disruptive innovations, build exponential revenues, and beat your competitors.” In other words, having IT bury its head in the sand is going to mean missing some opportunities that don’t come along very often to reshape the competitive landscape.

Especially when you couple the support of these tablets and other mobile devices with the changes coming about with the rise of cloud computing.

Look in the mirror

In the end, says Andi, “it’s all about you! ...The bottom line is that you — as an individual, as a consumer, as an employee, as an IT professional — are responsible for a radical change affecting business, government, and IT. You are both driving this change as a consumer of social, mobile, and cloud applications; and being driven by this change as an IT professional adapting to these new customer behaviors.”

Maybe TIME Magazine wasn’t wrong a few years back when they named You as their Person of the Year (congrats, by the way) with a big mirror-like thing on their front cover. It’s just that the revolution always takes longer than people think, and the results are never quite evenly distributed.

I’m a perfect example. I've been involved in cloud computing for many years, but didn’t join this particular part of the revolution -- the part where I expect flicking my fingers on a piece of glass will get me access to what I want -- until this past weekend.

But I’ll probably be confounding IT departments left and right from now on. Make it so.