Monthly Archives: February 2009
I am a big fan of Amazon’s Kindle. My husband gave me one as a gift when they first came out, and I went from taking it on the road to read the occasional thing as supplement to the paperbacks and magazines I was going through, to making it my sole form of reading material while out on travel, to wanting to read just about everything on it, period. The ability to change the font size, essentially allowing me to read every book as if it were in large print, is a big reason why — it’s easier to read bigger print when you’re in something that’s moving, as it creates less eyestrain.
I consume an enormous number of books (around a book a day if I’m traveling, and around half that if I’m not). Books are one of the most significant expenses in my household; my husband and I are both voracious consumers of fiction and non-fiction, and we mostly read different books. Kindle helps me spend a lot less on books, sort of — I pay less for the individual books, but because of the convenience, I also read even more than I normally would. And whereas I often used to wait for the paperback, now I buy books as soon as they come out in Kindle form. Plus, while business books are often grotesquely expensive for relatively limited value, especially when they’re in hardback, at Kindle prices, I don’t mind buying a book for the one cool idea in it, instead of standing around in the bookstore, flipping pages. Finally, rather than buying a ton of books that accumulate in piles and sometimes eventually disappear onto the shelves before I actually read them, I read every item I download onto my Kindle.
New York Times reviewer David Pogue understands the Kindle. But opinion columnist Roy Blount totally fails to get it, using the NYT megaphone to whine that the text-to-speech function potentially steals money from authors who would otherwise be able to sell audiobooks.
Seth Godin loves his Kindle. And he has a bunch of great suggestions for taking the Kindle service to the next level. Among other things, he points out that authors need to embrace these new models as a source for lots of new forms of revenue generation, rather than obtusely trying to cling to the way things are.
You can fear the future, or you can think different and embrace it. Devices like the Kindle open up a wealth of opportunities to authors who are willing to seize them.
A few days ago, an unexpected side-effect of some new code caused a major Gmail outage. Last year, a small bug triggered a series of cascading failures that resulted in a major Amazon outage. These are not the first cloud failures, nor will they be the last.
Cloud failures are as complex as the underlying software that powers them. No longer do you have isolated systems; you have complex, interwoven ecosystems, delicately orchestrated by a swarm of software programs. In presenting simplicity to the user, the cloud provider takes on the burden of dealing with that complexity themselves.
People sometimes say that these clouds aren’t built to enterprise standards. In one sense, they aren’t — most aren’t intended to meet enterprise requirements in terms of feature-set. In another sense, though, they are engineered to far exceed anything that the enterprise would ever think of attempting themselves. Massive-scale clouds are designed to never, ever, fail in a user-visible way. The fact that they do fail nonetheless should not be a surprise, given the potential for human error encoded in software. It is, in fact, surprising that they don’t visibly fail more often. Every day, within these clouds, a whole host of small errors that would be outages if they occurred within the enterprise — server hardware failures, storage failures, network failures, even some software failures — are handled invisibly by the back-end. Most of the time, the self-healing works the way it’s supposed to. Sometimes it doesn’t. The irony in both the Gmail outage and the S3 outage is that both appear to have been caused by the very software components that were actively trying to create resiliency.
To run infrastructure on a massive scale, you are utterly dependent upon automation. Automation, in turn, depends on software, and no matter how intensively you QA your software, you will have bugs. It is extremely hard to test complex multi-factor failures. There is nothing that indicates that either Google or Amazon are careless about their software development processes or their safeguards against failure. They undoubtedly hate failure as much as, and possibly more than, their customers do. Every failure means sleepless nights, painful internal post-mortems, lost revenue, angry partners, and embarrassing press. I believe that these companies do, in fact, diligently seek to seamlessly handle every error condition they can, and that they generally possess sufficient quantity and quality of engineering talent to do it well.
But the nature of the cloud — the one homogenous fabric — magnifies problems. Still, that’s not isolated to the cloud alone. Let’s not forget VMware’s license bug from last year. People who normally booted up their VMs at the beginning of the day were pretty much screwed. It took VMware the better part of a day to produce a patch — and their original announced timeframe was 36 hours. I’m not picking on VMware — certainly you could find yourself with a similar problem with any kind of widely deployed software that was vulnerable to a bug that caused it all to fail.
Enterprise-quality software produced the SQL Slammer worm, after all. In the cloud, we ain’t seen nothing yet…
CDNetworks buys Panther Express
For many months now, CDN industry insiders have gossiped that Panther Express was in financial trouble. Panther was caught with the bad luck of mistiming the funding cycle, leaving them to try to raise capital at a point when the capital markets were essentially frozen. Moreover, a large percentage of their revenues were tied to no-commit or limited-commit contracts, and with CDN prices in free-fall for much of 2008, Panther was doubly screwed from the perspective of the money guys. As time wore on, the likelihood of an acquisition by either a rival CDN or a carrier wanting to get into the space became more and more likely — but the longer the potential acquirers could wait to pull the trigger, the more cheaply they could buy the accompany, especially since rumors of Panther’s financial difficulties were starting to scare off potential customers.
Enter CDNetworks, a global CDN based in South Korea, who in the last year has been aggressively trying to penetrate the North American market. CDNetworks acquired Panther Express yesterday, in a deal structured so that it merged its US and European-based operations with Panther. Panther’s CEO Steve Liddell (who has experience working with Asian-based companies through his past experience as president of Level 3’s Asia business) will lead the new entity.
Dan Rayburn has offered some numbers and claims the acquisition values Panther at about $5 million — which would be about one-quarter of its 2008 trailing revenues, and would leave me wondering how that compares to the book value of Panther’s deployed eqiupment.
I’ll simply say that, although I agree with Dan that this acquisition basically has zero impact on other players or on pricing in the market, I have a very different perspective on the acquisition itself, and carrier opinion of this space, and of the general market opportunity (especially in the context that CDN is much more than video), than Dan does. Gartner clients who want to talk about it, you’re welcome to schedule an inquiry with me.
(Sorry. I started to write a long and detailed analysis, and then realized that I was crossing the line on what Gartner views as acceptable analyst blogging, and what is full-fledged analysis that ought to be reserved for paying clients.)
Interesting tidbits for the week
A bit of a link round-up…
My colleague Daryl Plummer has posted his rebuttal in our ongoing debate over cloud infrastructure commoditization. I agree with his assertion that over the long term, the bigger growth stories will be the value-added providers and not the pure-play cloud infrastructure guys, but I also stick to my guns in believing that customer service is a differentiator and we’ll have a lot of pure-plays, not a half-dozen monolithic mega-infrastructure-providers.
Michael Topalovich, of Delivered Innovation, has blogged a post-mortem on Coghead. It’s a well-written and detailed dissection of what went wrong, from the perspective of a former Coghead partner. Anyone who runs or uses a platform as a service would be well served to read it, as there are plenty of excellent lessons to be learned.
Richard Jones, of Last.fm, has put up an annotated short-list of distributed key-value stores (mostly in the form of distributed hash tables). He’s looking for a premise-based rather than cloud-based solution, but his commentary is thoughtful and the comments thread is interesting as well.
Also, I have a new research note out (Gartner clients only), in collaboration with my colleague Ted Chamberlin: evaluation criteria for Web hosting (including cloud infrastructure services in that context), which is the decision framework that supports the the Magic Quadrant that we’re anticipating publishing in April. (Also coming soon, a “how to choose a cloud infrastructure provider” note and accompanying toolkit.)
Single-function clouds are unlikely
GigaOm has an interesting post on HP’s cloud vision, which asserts that HP’s view of the future is that service providers will reducing complexity by delivering only one application (scaling up their own infrastructure in a monolithic way), and that generalized infrastructure-as-a-service (IaaS) providers will not be able to scale up in a profitable manner.
Setting the specifics of what HP does or does not believe aside, I think that it’s highly-unlikely that we’ll see super-specialization in the cloud. There are, of course, software vendors today who make highly specialized components, that in turn are incorporated into the software of vendors further up the stack — today, those are ISVs that sell libraries, Web 2.0 companies with mashable components, and so forth. But as software companies get more ambitious, the scope of their software tends to broaden, too. In the future, they may want to become the masher rather than the mashed, so to speak. And then they start wanting to become diversified.
For an example, look at Oracle. Originally a database company, they now have a hugely diversified base of enterprise software. Why should we believe that a cloud-based software company would be any less ambitious?
Certainly, it is more difficult and more expensive to manage general-purpose compute than it is to manage specific-purpose compute. But there’s a great deal more to driving profitability than keeping costs down. Broader integration has a business value, and the increase in value (and the price the customer is willing to pay) can readily outpace the increased infrastructure cost.
To take another example, Google runs incredibly efficient single-purpose compute in the form of their search farms. Yet, they are trying to broaden their business to other services, both for the potential synergies, and because it is incredibly dangerous for a business to be too narrow, since that limits its growth and vastly increases its vulnerability to any shifts in the market.
I don’t think successful software companies will confine themselves to delivering single applications as a service. And I think that IaaS providers will find cost-effective ways to deliver appropriate infrastructure to those SaaS companies.
I have found at a partial solution to my communication proliferation problem. The tool I’m employing, at least for the moment, is Digsby, a free client that combines cross-platform instant messaging with access to social networking sites like Facebook, LinkedIn, and Twitter. This has replaced my usual IM client (Trillian, which I like a lot), but since I exchange very few IMs, that’s not an issue. However, the aggregated volume of communication still feels like it’s too much. I need robust filtering capabilities, and I have the feeling that I may need to write my own client (or hack an open-source client) in order to get that.
I’ve recently read Pete Blackshaw’s Satisfied Customers Tell Three Friends, Angry Customers Tell 3000, which is a well-written, methodical introduction to consumer-generated media (CGM, also known as UGC, user-generated content). I’d recommend the book to anyone who hasn’t read a book on the topic; if you’re social-media savvy, chances are you won’t learn much (if anything) new, but the anecdotes are entertaining and useful, and the structured approach provides good framework language.
Thus, trust, credibility, and authenticity in corporate engagement are very much on my mind, at a time when there’s a new (resurfaced) controversy regarding local-review site Yelp, which is being accused of manipulating user reviews to gain advertising revenues. Naturally, Yelp denies any extortion of local businesses.
As an analyst, I belong to an industry which is constantly being questioned about the credibility and authenticity of its commentary — the age-old question of whether it’s a “pay to play” business where vendor clients receive ratings and recommendations that are more favorable than those that non-clients get. I still find myself having to stress to clients and non-clients alike that Gartner opinion cannot be bought. It’s one of the genuinely great aspects of working here — the organizational commitment to integrity. This is not to say that there aren’t conflicts — a vendor client has more avenues with which to express their unhappiness with an analyst’s opinion, and attempt to influence it in a more positive direction. But in the end, we pride ourselves on serving our IT buyer clients with honest advice — which means that vendor dollars can’t be allowed to influence analyst opinion.
I imagine that for any organization which provides reviews and recommendations as part of its business, and which accepts money from the entities being rated, has problems with rogue salespeople who attempt to imply, or even outright state, that paying for services means more favorable positioning. So the question is, what’s the organization’s attitude, from the CEO down, towards these things? Is it a wink-wink nudge-nudge thing where the organization only pays lip service to neutrality, is it a don’t-ask don’t-tell thing where the organization is willing to turn a blind eye as long as it doesn’t cause obvious problems, or is the organization really dedicated to ensuring that dollars don’t alter anything?
Which of these categories does Yelp fall into? I’m a pretty engaged consumer — I read reviews on Yelp, and I write them from time to time. I’ve got a keen interest in knowing.
Application delivery network adoption
A long-standing puzzle for myself and my various colleagues who cover application-fluent networking: Why don’t more SaaS providers adopt application delivery networks (ADNs), either via a service, or via application delivery controller (ADC) hardware?
Even if a SaaS vendor perceives their performance as being just fine for the typical US-based user, performance is often an issue in Europe, and frequently deteriorates sharply in Asia, especially China, and is erratic everywhere else depending on the quality of the country’s connectivity. (Change the names of the regions if the data center isn’t in North America.) Deploying an ADN helps to bolster performance for these users. And if it’s not cost-effective to do that for all users, why not charge extra for an accelerated service? (Yes, we understand that there are issues like “if we offer an accelerated service, are we implying our regular service is slow?” but really, that’s just a marketing issue. Performance can be a competitive differentiator and it’s also a revenue opportunity.)
Two interesting recent examples:
- Fog Creek’s CoPilot gets Akamai IP Application Accelerator service
- F5 does a China solution for its DevCentral portal
Yes, times are tough right now, so a SaaS company does have to evaluate the ROI carefully, but any SaaS provider with performance issues owes it to themselves to give this stuff a look. (And SaaS customers who have performance issues ought to be poking at their providers.)
Discipline and agility are not opposites
Too many service providers (and companies in general) use “discipline” as an excuse for “lack of agility”. Discipline does not mean appointing a committee to study the problem for the next year. Exercising caution and prudence does not mean failing to act. Laying a solid foundation does not mean standing around doing the equivalent of watching the concrete set. This misguided notion of discipline is made even worse if the committee sits around drawing personal conjectures based on fear, and concluding that moving with paralyzing slowness (because, I guess, sudden motion draws predators) is the only safe possibility.
There are highly agile companies out there who study and solve problems in a rigorous way — they go out and gather data, they analyze the data, they come to a conclusion, they come up with a solution, they decide what they want to measure to determine the success or failure of the solution, and go out and act. There’s enormous value in swift, decisive, fact-based action. Bonus points if the decision-making is focused on what delivers value to the customer, and that value (whether qualitative or quantitative) can be clearly articulated and measured.
Does cloud infrastructure become commoditized?
My colleague Daryl Plummer has mused upon the future of cloud in a blog post titled “Cloud Infrastructure: The Next Fat Dumb and Happy Pipe?” In it, he posits that cloud infrastructure will commoditize, that in 5-7 years the market will only support a handful of huge players, and that value-adds are necessary in order to stay in the game.
I both agree and disagree with him. I believe that cloud infrastructure will not be purely a commodity market, specifically because everyone in this market will offer value-added differentiation, and that even a decade out, we’ll still have lots of vendors, many of them small, in this game. Here’s a quick take on a couple of reasons why:
There are diminishing returns on the cost-efficiency of scale. There is a limit to how cheap a compute cycle can get. The bigger you are, the less you’ll pay for hardware, but in the end, even semiconductor companies have to make a little margin. And the bigger you are, the more you can leverage your engineers, especially your software tools guys — but it’s also possible that a tools vendor will deliver similar cost efficiencies to the smaller players (think about the role of Virtuozzo and cPanel in shared hosting). Broadly, smaller players pay more for things and may not leverage their resources as thoroughly, but they also have less overhead. It’s important to reach sufficient scale, but it’s not necessarily beneficial to be as large as possible.
This is a service. People matter. It’s hard to really commoditize a service, because people are a big wildcard. Buyers will care about customer service. Computing infrastructure is too mission-critical not to. The nuances of account management and customer support will differentiate companies, and smaller, more agile, more service-focused companies will compete successfully with giants.
The infrastructure itself is not the whole of the service. While there will be people out there who just buy server instances with a credit card, they are generally, either implicitly or explicitly, buying a constellation of stuff around that. At the most basic level, that’s customer support, and the management portal and tools, service level agreements, and actual operational quality — all things which can be meaningfully differentiated. And you can obviously go well beyond that point. (Daryl mentions OpSource competing with Amazon/IBM/Microsoft for the same cloud infrastructure dollar — but it doesn’t, really, because those monoliths are not going to learn the specifics of your SaaS app, right down to providing end-user help-desk support, like OpSource does. Cloud infrastructure is a means to an end, not an end unto itself.)
It takes time for technology to mature. Five years from now, we’ll still have stark differences in the way that cloud infrastructure services are implemented, and those differences will manifest themselves in customer-visible ways. And the application platforms will take even longer to mature (and by their nature, promote differentiation and vendor lock-in).
By the way, my latest research note, “Save Money Now With Hosted and ‘Cloud’ Infrastructure” (Gartner clients only) is a tutorial for IT managers, focused on how to choose the right type of cloud service for the application that you want to deploy. All clouds are not created equal, especially now.