GigaOM » How-to (hack, pack, & backpack)

Intel’s Multi-Billion-Dollar Cost of Doing Business

Stacey Higginbotham — Tue, 19 Oct 2010 22:31:55 +0000

Intel’s patriotic investment in American manufacturing news blast worked so well last year when it made a fairly big to-do over its planned capital expenditures that it’s taken a similar tactic this year, showcasing its $6 billion to $8 billion in planned manufacturing investments.

Apple TV: 250,000 Sold In Six Weeks

Ryan Lawler — Mon, 18 Oct 2010 22:15:45 +0000

Apple didn’t announce the number of Apple TVs sold as part of its earnings release, but Steve Jobs was on the company’s investor call today and gave the first official numbers for how many of the brand new set-top box have seen purchased. So far, the sales are pretty good: Jobs reported that the company has sold more than 250,000 Apple TVs since the company first made the product available in early September.

Apple first announced the new Apple TV broadband set-top box on September 1, pricing it at $99 and announcing a new business model for streaming rentals of TV episodes and movie files. With the new streaming model, Apple rents TV episodes from Fox and ABC for 99 cents, compared to the $1.99 price for standard definition episode downloads and HD downloads for $2.99.

To put that into perspective: Roku reported at the beginning of this year that it had sold 500,000 of its set-top devices after about a year-and-a-half on the market, and hopes to have sold more than a million of its box by the end of the year. So Apple is tracking well ahead of other, similar devices already in the market.

That said, Apple TV is still very much a hobby when compared to the other products it makes. Apple sold 4.2 million iPads and 3.9 million Macs in the third quarter, as well as 9 million iPods and 14 million iPhones. While those products were available all quarter, compared to just the last several weeks like the Apple TV, it’s clear that it could be some time before Apple’s video set-top box becomes a significant portion of Apple’s business.

Related content on GigaOM Pro: (subscription required)

Ray Ozzie to Leave Microsoft: Has The Future Left the Building?

Om Malik — Mon, 18 Oct 2010 21:29:57 +0000

Ray Ozzie, the chief software architect with Microsoft is leaving the Redmond, Wash.-based software giant. Ozzie, is widely viewed as someone who tried to change Microsoft’s internal attitudes towards cloud computing. He is credited for Microsoft’s move into the cloud with its Azure efforts.

Hey Shareholders, Capex Means Cash in the Cloud!

Derrick Harris — Fri, 15 Oct 2010 23:27:42 +0000

Om’s post about Google’s soaring infrastructure spending got me thinking about the hypocrisy in the way we assess web companies’ decisions to splurge on the very equipment that makes them tick. Startups are either expected to or praised for spending on more infrastructure, while public companies feel the wrath of financial analysts when they do the same.

Users come to love services like Facebook and Twitter (even Foursquare) so much that it resembles a national crisis among some circles when one of their sites goes down for a few hours. How do they try to avoid these occurrences in the future? Well, they spend some of those millions improving their physical infrastructures and creating specialized software to address unique needs. Greenpeace aside, I don’t recall hearing any complaints about Facebook’s first data center, nor about Twitter’s planned data center, and these companies aren’t even making money.

So why is it that financial analysts and shareholders get so up in arms when massive public web companies like Amazon or Google embark on spending kicks? After all, they bring in billions in revenue each year, and Google is responsible for more web traffic than Facebook and Twitter combined. Because it detracts from near-term profitability, that’s why. Om calls such analysts “idiots and short-term thinkers,” and he’s spot-on with his analysis of the situation.

Imagine if Google’s search engine, or its paid Google Apps service, went down as frequently as Twitter. What about Amazon Web Services, which hosts a good number of popular web sites and relatively important enterprise applications? What if these companies never rolled out new services because that would require spending more money on infrastructure? The answer is that they wouldn’t be too popular for too long. Companies delivering services via the web have to spend money – on infrastructure – to make money.

I can only imagine the number of cloud computing providers who’d love to have the problem of needing to scale to meet demand, and actually having the cash to do so. That day will never come if they don’t do infrastructure right in the first place.

When we’re talking about companies for which millions today might mean billions tomorrow, it might behoove shareholders and analysts to lighten up a little bit on the bottom line.

Related content from GigaOM Pro (sub req’d):

Oct. 15: What We’re Reading About Infrastructure

Derrick Harris — Fri, 15 Oct 2010 21:14:34 +0000

Bolstering the argument that money makes the world go ’round, Google’s earnings, and spending, were big news today, as is the third-quarter decrease in VC funding. AMD’s “better than expected” net loss also garnered attention, and we revisit the notion that cloud computing will take away individuals’ money in the form of job losses.

I, Cloud (From High Scalability) I agree that cloud computing probably will result in more job-shifting than job-cutting, but I wouldn’t be so sure that software can’t use context to make human-like judgment calls. (see, e.g., IBM’s Watson supercomputer).

Google Earnings: “Excellent” Quarter; Momentum of Display, Mobile Highlighted (From ZDNet) Google is a moneymaking machine, plain and simple. Even with enormous infrastructure spending, it’s still raking in more than $2 billion in quarterly profit.

Verizon Medical Data Cloud on Track to Serve 350,000 Doctors (From SearchCloudComputing.com) It looks like the “Field of Dreams” logic works for data clouds, too: If you build it, they will come. Even one of three doctors uploading records there could mean a big improvement in health care efficiency.

Notes on Data Warehouse Appliance Prices (From DBMS2) Just in case you were wondering how much one of the new breed of data warehouse appliances actually cost. The answer can be complex, and discounts are plentiful.

Opteron 6100 Server Ramp Underway, Says AMD (From The Register) AMD could use some good news considering that its net loss was considered “better than expected.” Can its Opteron 6100s make their way into enough servers — which are selling like crazy — to spur a rebound?

For more cloud-related news analysis and research, visit GigaOM Pro.

Image courtesy of Flickr user liewcf.

New Relic Gets Another $10M, Proves SaaS Profitability

Derrick Harris — Thu, 14 Oct 2010 15:28:34 +0000

SaaS startup New Relic has received an additional $10 million in venture capital for its application performance management (APM) offering that targets both data centers and the cloud. Tenaya Capital and Allen & Co. led the Series C round, with existing investors Benchmark Capital and Trinity Ventures also chipping in. The round brings New Relic’s total investment to $20 million, proving the importance of management and monitoring as companies adopt cloud computing, as well as the power of the SaaS 2.0 model.

What separates New Relic from other APM providers is that its RPM service monitors live software from the inside instead of simply monitoring external web application metrics. RPM comes in five editions, ranging from the free Lite version, which gives basic monitoring capabilities, up to the Enterprise version, which provides a range of capabilities to discover and cure performance issues. According to founder Lew Cirne, the internal view helps customers identify problems early and avoid “the thorniest issues,” citing the recent Foursquare outage as a problem that might have been avoided if the company could have been proactive in addressing the problem.

Formerly, New Relic supported only Java and Ruby web applications, but as of today, it also supports .NET and PHP applications. This is nothing new for cloud-based services, which tend to start with a focused offering and expand their customer bases as they grow by adding additional language support. Even larger companies like Google and Salesforce.com took this approach with their cloud offerings. More applications mean more money, after all.

New Relic’s already-expansive footprint would seem to underscore the value of its service and of the SaaS model. The company counts just about every cloud provider as a technology partner (Joyent and Heroku offer it as an add-on in their offerings), and the company has experienced 200 percent growth annually since launching in 2008. Cirne says the company presently has 5,600 customers running the service in production, with 900 of them paying New Relic directly. Among them are large enterprises running New Relic within their data centers.

Perhaps more important than the product, however, is how the economics of New Relic underscore the power of the SaaS model –- specifically, the freemium model prevalent among SaaS 2.0 vendors. Nine hundred paying customers is no small feat when you consider that the company just increased its sales force to three from two a couple of weeks ago. Cirne contrasts New Relic’s approach to that of Salesforce.com, which customers can’t start using without speaking with multiple salespeople, as well as to the last company Cirne founded, Wily (an on-premise APM vendor), which he sold to CA for $375 million.

When CA bought Wily, Cirne says it had 500 customers, 270 employees (about 150 of them in sales) and had raised $45 million. Keep in mind, New Relic just closed its latest round. Most of what it has accomplished was with about $10 million. Cirne says $20 million is way more than New Relic needs to achieve profitability, but it’s good to have some extra cash to capture opportunities in what he thinks will be a billion-dollar business.

Image courtesy of Wikipedia Commons user KlausFoehl.

Related Research about on Big Data from GigaOM Pro:

Credit Suisse Spawn DynamicOps Enters Private Cloud Game

Derrick Harris — Tue, 12 Oct 2010 01:40:27 +0000

The cloud-management space got a little more crowded today with the release of DynamicOps’ Cloud Automation Center. That market is currently full of hot startup vendors like Cloud.com, Nimbula, Eucalyptus and Abiquo, all of whom are trying to give companies out-of-the-box internal clouds that mimic the public-cloud experience while maintaining existing IT policies, so DynamicOps has a great deal of competition. However, the Lexington, Mass.-based company does have a couple of aces in the hole that might make it particularly appealing to enterprise customers.

One is its evolution from an internally developed virtualization-management solution at megabank Credit Suisse into, a spinoff company selling that software to the greater IT community in 2008. Credit Suisse was managing thousands of virtual servers and desktops while DynamicOps’ underlying Virtual Resource Manager software was developed, so scalability isn’t an issue. In fact, said VP of Marketing Rich Bordeaux, one customer currently manages 30,000 VMs and virtual desktops and is looking to have more than 60,000 within 18 months.

Furthermore, because the bank wasn’t about to move certain application from physical servers, physical-resource management is inherent in the Cloud Automation Manager. The new features in Cloud Automation Manager are what make it truly cloudy, though: multitenancy, self-service provisioning, and the ability to provision and manage resources from Amazon EC2.

DynamicOps’ other ace is that Cloud Automation Manager is an integral part of Dell’s recently announced Virtual Integrated System (VIS) offering. The DynamicOps software is the foundation of the VIS Self-Service Creator component. VIS is Dell’s attempt to combat competitors’ converged infrastructure solutions (e.g., Cisco UCS and HP BladeMatrix), so such an OEM deal should help validate DynamicOps’ claims of enterprise-readiness (if the Credit Suisse heritage doesn’t already do so).

Of course, it’s hard to get too excited about any cloud software right now before organizations actually start buying it. Web hosts and MSPs have been buying up software from vendors like VMware, Cloud.com and 3tera, but most other businesses appear to be doing no more than dipping their toes in the water right now. DynamicOps has a strong virtualization-management history, though, and an existing customer base to tout (and sell), so perhaps it can break in where others appear to be left outside.

Related Research about NoSQL Databases from GigaOM Pro:

Come Talk With Us About the Open-Source Enterprise

Michael Wolf — Wed, 22 Sep 2010 21:52:47 +0000

The promise of a completely open-source enterprise is near, but the nagging question remains: Will it work as a business for the technology industry? On Sept. 29, about 75 entrepreneurs, executives and investors will gather at the GigaOM San Francisco offices to discuss that very question.

Exalogic: Larry Gets the Cloud Now & He Wants It All

Gary Orenstein — Mon, 20 Sep 2010 22:20:25 +0000

Oracle today announced a slew of new products that, at least on paper, deliver one of the most comprehensive enterprise computing solutions around. The company has extended the concept of “stack” further than any other technology player, by covering the hardware from compute to storage, and every piece of software in between. While competitors like HP and potentially IBM, Dell, or Cisco might one day get there, Oracle appears far ahead of the other big technology vendors aiming to deliver the next-generation infrastructure for the data center. Currently, many of the hardware players will still have to turn to VMware if they can’t do the software themselves. Oracle has built and bought what it needs to go it alone.

More specifically, Oracle announced the Exalogic Elastic Cloud, a tightly coupled hardware and software appliance that packs the most powerful CPUs, gobs of DRAM, InfiniBand networking, solid state drives, and high-capacity SATA drives into a server. The configurations come in quarter-, half-, and full-rack configurations and can scale up to eight racks total. The specifications are indeed impressive, if you like the expensive hardware approach. Make no mistake; this is not the direction of scale-out, commodity hardware that many people refer to around cloud computing.

Instead, Oracle has positioned Exalogic as the one stop shop to provision virtual machines, middleware, and applications, to offer a tightly integrated hardware and software solution that will be beneficial for a large number of companies that don’t want to assemble their own infrastructure layers. Essentially, this is Oracle’s build-your-own private cloud for the Fortune 1000 set.

But Exalogic is just one piece of the overall puzzle. Oracle intends to couple many other pieces that should, in theory, provide a cloud-like infrastructure within a corporation. This includes its own version of Linux, dubbed Unbreakable Linux, an option that breaks away from strict Red Hat compatibility in order to take advantage of complimentary hardware and software capabilities, or a version of Oracle Solaris. Additionally, Oracle is touting Exalogic integration with WebLogic, its Java application platform, and Coherence, its distributed data manager. And of course, there’s Exadata, the precursor to Exalogic, which supports high-capacity, data-intensive workloads.

So what does this dizzying array of announcements mean? Here are my takeaways:

There’s still big business in integration. Plenty of customers have no interest or intent to qualify different hardware components, test for compatibility, manage configurations and optimize for performance. A pre-configured solution makes those headaches go away. Yes, cost is a factor (integration doesn’t come cheap), but obviously, not for everyone.

Java will be the money-maker for cloud computing over the next few years. There’s a reason VMware focuses on Java, that Google App Engine added Java, and that Oracle now emphasizes Java performance. Enterprises pay big bucks for Java speed, and those moving to cloud architectures want as much Java performance as they can get.

The proverbial “you can move your apps” card must be there. Similar to VMware and its emphasis on application portability to VMforce or AppEngine, Oracle is also making application portability possible to Amazon EC2. Oracle plans to create specific Amazon Machine Instances for Oracle apps.

More tightly coupled solutions simplify an increasingly complex stack. With architects now needing to plan for compute, middleware, data processing and analytics all at the same time, the complexity can be overwhelming. Oracle aims to simplify that by providing a set of pre-configured options for those that want to focus more on their business and less on their technology.

Specs still sell. Oracle’s news announcements about Exalogic and Exadata read like a sports car fanatic’s wish list. The underlying components still matter to many, and Oracle has spared no expense to make sure they include all the bells and whistles.

All in all, the payoff from the hardware and expertise acquired from Sun appears to be showing. Oracle has amassed a comprehensive stack that covers the major software layers, plus an impressive array of server and storage power. Oracle’s expertise in middleware shouldn’t be discounted as the race to improve the various Platforms-as-a-Service continues.

Perhaps the biggest challenge for Oracle, and as I also mentioned, for VMware following VMworld, will be how to continue sales momentum with a product line that includes much more than traditional Oracle software markets. But don’t discount Larry Ellison in any of this. Though he mocked the cloud-washing early on, his company now has a comprehensive cloud portfolio. He knew hardware would be needed to compete with HP, IBM, Dell, and Cisco; he got a snazzy portfolio on the cheap in Sun; and he knows how to pull in top executive talent to keep the machine moving forward. This game is far from over.

Gary Orenstein is the Host of The Cloud Computing Show.

Image via Flickr user yatra.

Related GigaOM Pro research (sub req’d): For Open Cloud Computing Look Inside Your Data Center

NoSQL Databases Providing Extreme Scale and Flexibility

For Greener Data Centers, Give Servers An Oil Bath

Katie Fehrenbacher — Fri, 03 Sep 2010 16:20:35 +0000

Dunking servers in a bath of oil sounds like the fastest way to break some very expensive hardware. But not for startup Green Revolution Cooling, which builds energy efficient liquid-cooled servers and its first customer collocation firm Midas Networks will implement the technology later this year.