GigaOM » Tech

The biggest market you’ve never heard of

Ed Robinson — Sun, 19 Jun 2011 19:00:39 +0000

The tech bubble of the late ’90’s was fueled largely by the promise of universal high-speed Internet access. As millions of consumers gained access to the Internet, new market opportunities emerged. But today, content is so heavy, and networks so overburdened, that more efficient use of the network is a critical behavior.

The state of web content today

As richer, more dynamic, more interactive sites have hit the Web; the existing infrastructure has become insufficient. While high-speed broadband has tried to meet the infrastructure demands of the exploding volume and size of content on the Web, it’s clear that throwing pure infrastructure at the problem isn’t enough.

Two new markets emerged from these challenges: the content delivery network market (CDN) and the application delivery controller market (ADC). Put simply: These are technologies that help make your experience on the web a lot faster, while still using the same infrastructure that has been in place for the past two decades.

Remarkably, those two markets are now struggling to keep up with the explosive growth of the web. Sites are too big, too dynamic, and too rich for our existing infrastructure, and prevailing techniques for optimizing performance.

Today, we’re embarking upon the third major evolution in modern web performance. Web content optimization and acceleration is one of the largest market opportunities in the tech sector today, and it’s going to pave the way for the next major era of the Internet. Without it, innovation gets throttled.

The technologies we’re currently using to speed up the web need to be supercharged. They need extra help. That’s where making sure web content is efficiently delivered comes in. It’s about the conservation of bandwidth and the compression of megabytes, especially on mobile networks.

Four problems you can’t ignore

Most of the web performance challenges we face today can be traced to four basic trends:

Third-party content. Any given web site incorporates vast amounts of third-party content. This includes content such as advertisements, widgets and syndicated feeds.
Dynamic. Sites are now required to be more dynamic than our infrastructure can handle. Twitter feeds are constantly changing, and the data can’t be cached, and furthermore, we expect a high degree of personalization and individually relevant experiences when we visit sites.
More, more, more. We’re experiencing a content explosion: Sites have more pages, more pictures and more videos packed into the pages than ever before.
New devices. Myriad new devices hit the market every month, all of which are Internet-enabled. This doesn’t just mean more laptops and iPads; we’re also talking about refrigerators, low-cost home security cameras, and even cars!

And why is this happening? There’s a new party in town, and its called social media. Our problems aren’t capacity problems; it’s just that our content delivery infrastructure wasn’t designed for what’s happening. The existing infrastructure is designed on three premises and assumptions:

Single origin. Most content will originate from the same web servers, so if these are working properly, then everything is good.
Static content. Most information will stay the same, and therefore can be cached across the data center and Internet.
Fast delivery. Because content comes from a single origin and doesn’t change, caching and route optimization can deliver everything quickly.

Social media turns these assumptions on their heads. Content is mashed-up, syndicated, streamed from everywhere — with different qualities of service. So even if you’re paying $500,000 for traditional performance solutions, your pages will still slow down to the lowest common denominator, such as a slow ad service or the slower speeds of a streamed page from Facebook.

Yes, we can build new infrastructure, but it will take too long, and it may not be enough. We can throw more of the same performance technology at it, but this only helps so much, and the traditional technology doesn’t do anything for today’s dynamic content, which can’t be cached. At the end of the day, these four factors have driven intense demand for a new type of web acceleration.

With our powers combined, we are …

The good news is that we have the technology to solve the problem, and there has already been a good deal of investment to put the wheels in motion. We’re seeing the many web performance players converge to do this.

Recently, Limelight Networks acquired AcceloWeb for up to a rumored $20 million in a cash and stock deal. AcceloWeb’s technology does precisely what I’ve hinted at so far: It accelerates web content so that it can travel faster over our existing Internet infrastructure. Limelight, a traditional CDN company, is making a large investment in Web content optimization and acceleration. These are two fundamentally different markets converging under one company, yet we hardly heard any talk about the strategy behind the investment.

Similarly, Google recently announced that Google Analytics now offers a Site Speed Analytics Report. It was greeted with applause from the web performance community, but nobody really heard about how this “feature” had much broader implications for the web.

Google isn’t just helping you measure your site’s speed; they want the Web to be lightning fast. It’s critical to the future of their business that the web isn’t crippled by performance woes.

Their revenue is still largely ad-based, which contribute costly seconds to load times if we don’t find a solution. Not to mention: The faster a site loads, the more ads Google can serve. Google cares about web performance because it’s absolutely critical to their business and the future of the web itself.

Just how big is this?

This isn’t just the market opportunity for web content acceleration that’s exciting here. What’s more important is the future of the web, and what this evolution in web performance will spawn.

We’re talking about webscale personalization that isn’t held back by performance problems. Personalization is the web topic du jour, but we’re not going to reach the promise of true web personalization if we can’t load web pages faster than we’re doing on average today.

Similarly, the mobile web is going to face major obstacles if we can’t tune our apps to perform on even the most troubled networks. And we sure as heck aren’t going to usher in the future of virtual personal assistants if we can’t conduct complex processing and deliver that content at the speeds that consumers demand.

Oddly enough, the success or failure of these sexy technologies hinges on a critical evolution in web performance. All of a sudden, the emerging web acceleration and content optimization market is starting to look a lot sexier to investors, entrepreneurs and incumbent technology companies alike.

Ed Robinson is the CEO of Aptimize, a company that produces software to accelerate websites.

Related research and analysis from GigaOM Pro:
Subscriber content. Sign up for a free trial.

Limelight, Looking to Add Value, Buys Ad Startup Kiptronic

Liz Gannes — Thu, 21 May 2009 20:43:41 +0000

As CDNs try to evade commodification, they’re looking outward to bring more value-added services to their clients. To that end, Limelight Networks today announced its first-ever acquisition: dynamic video and audio ad insertion startup Kiptronic, which to date had raised $9 million in funding. Dan Rayburn reports that the value of the deal is around $12 million, but while Limelight says that’s inaccurate it’s declined to offer up an alternate figure.

Last year Akamai also bought an ad startup — aCerno, for behavioral advertising. Kiptronic handles targeting, too, but its focus is dynamically delivered ads to media in any environment — on a mobile device, on consoles, online or offline. It’s possible this could turn into a play for Limelight to use its precise content delivery information to power localized advertising on any device.

Related research and analysis from GigaOM Pro:
Subscriber content. Sign up for a free trial.

AT&T Creates Private CDN for Corporate Video

Stacey Higginbotham — Tue, 21 Apr 2009 14:34:28 +0000

Starting today, enterprise customers can get all the bandwidth for high-definition video they need from AT&T, which has launched a private content delivery service for video inside company firewalls. The explosion of video inside corporate networks is straining resources, according to the carrier. But unlike the tiered service that AT&T is experimenting with for its last-mile consumer networks, it’s offering enterprise customers a service that helps them track, compress and prioritize video traffic within the network.

AT&T explains pretty clearly why this is necessary in its release:

“On the average business day, about one-third of the more than 17 petabytes of traffic traversing AT&T’s global backbone network is video content. A mere three years ago, video content traversing the AT&T network produced barely a blip,” said Roman Pacewicz, senior vice president of strategy and application services, AT&T Business Solutions.

AT&T, rather than become a dumb pipe that acts solely as a conduit for the video deluge, is hoping to monetize that video traffic. On the enterprise side, it’s offering products such as its AT&T Private Content Distribution Service for inside the firewall; it launched a content delivery network last summer that will likely tie into this offering by delivering content from outside, too. Other carriers such as Verizon, and even Internet backbone providers such as Level 3, have CDN efforts as well.

AT&T’s Private Content Distribution Service sounds like it may be based on, or is competitive to, Cisco’s media:net, which manages video within the corporation. That means Cisco has either scored a great client for its initial forays into managing corporate video, or its competition just got tougher, as many large enterprises would rather outsource such efforts than build their own.

Related research and analysis from GigaOM Pro:
Subscriber content. Sign up for a free trial.

Voxel Adds Cloud Computing With SilverLining

Stacey Higginbotham — Mon, 16 Mar 2009 14:03:55 +0000

Voxel, the managed hosting provider, has built its own cloud computing product called SilverLining that will compete with Amazon’s Elastic Compute Cloud and Rackspace’s CloudServers products. Thanks to its managed hosting business, Voxel plans to offer the same hybrid strategy that both Rackspace and ServePath offer, which combines dedicated servers with cloud computing for spikes in traffic or one-off projects.

Voxel adds another piece to these offerings with its content delivery network, says Raj Dutt, CEO and founder of the 10-year-old Voxel. Unlike Rackspace, which outsources CDN services to Limelight, Voxel can provide everything in house — and generates about $1 million in sales a month doing so. It also doesn’t make a distinction between buying bandwidth for pages delivered from the server and buying bandwidth to deliver content hosted on the CDN — a terabyte is a terabyte says Dutt.

The strategy of delivering bandwidth as just another component of the cloud, rather than a higher-margin add-on service will help commoditize the CDN business further. It’s a service many cloud operators are trying to provide, either on their own or through partnerships. Late last year Amazon launched its own CDN called CloudFront with competitive — and transparent — pricing.

“Every single CDN competes on price, and there are too many providers,” Dutt says. “And they’re all going after the same business, so what used to be high margin is now marginally better than selling IP transit.”

Last month, two CDN companies merged in a fire sale for one of the companies, signaling further stress in the industry. The top company in this space, Akamai, has been branching out into services and advertising. As cloud providers move to package CDNs with commodity compute cycles, we’ll see more CDN companies consolidate or get snapped up by cloud vendors aiming to bring them in house.

Related research and analysis from GigaOM Pro:
Subscriber content. Sign up for a free trial.

Amazon's CloudFront Could Storm Rival CDNs

Stacey Higginbotham — Tue, 18 Nov 2008 13:31:32 +0000

Today Amazon Web Services launched the beta version of its content delivery network service called CloudFront. As Om mentioned in September when the service was announced, this is a good move for Amazon, and something that may put the hurt on fellow CDNs such as Limelight and Akamai. Amazon will charge a usage-based fee, rather than a long-term contract, bringing CDN prices even lower for smaller web players who don’t have the scale to negotiate lower prices. Here’s how it works from the release:

The service caches copies of content close to end users for low latency delivery, while also providing fast, sustained data transfer rates needed to deliver popular objects to end users at scale. CloudFront works seamlessly with Amazon S3, where users store the original versions of objects delivered through the service. Customers need only put their objects into an Amazon S3 bucket and then register that bucket with the new service using a simple API call, which then returns a domain name used to access content through the network of edge locations.

Werner Vogels, CTO of Amazon, explains all about CloudFront on his blog.

A content delivery service that would extend Amazon S3 has been something that is very high on the wish list of our customers. They were already successfully using Amazon S3 for some of their content distribution needs, but many wanted the choice to do so with even lower latency and with higher data transfer rates to any place in the world.

He goes on to explain:

Using Amazon CloudFront is dead simple. Many of our private beta customers have reported that it only took them 10-15 minutes from the moment that they first signed up for the service to the moment that Amazon CloudFront was distributing their content.

The second Amazon Web Services principle that sets Amazon CloudFront apart is that no upfront commitments are necessary and you only pay for what you have used. There are no upfront fees or high volume requirements and no negotiations are necessary because we have published low prices from the start.

The second point is the more disruptive one. When Amazon announced its CDN in September we wrote,

Akamai is less likely to be impacted in the near term, but it further commoditizes the CDN business and forces a big shakeout in the industry, taking down the small and the weak. Akamai has been focusing on value-add services, as a way to stay ahead of the commoditization of the basic CDN services.

With prices ranging from 17 cents per gigabyte for the first 10 terabytes sent out a month, to 9 cents per GB for everything over 150 TB, the service seems to undercut the pricing offered by other CDNs for small to medium sized customers. It might be a good thing that Akamai’s looking at diversifying into online advertising.

Related research and analysis from GigaOM Pro:
Subscriber content. Sign up for a free trial.

FastSoft Tweaks TCP to Speed the Internet

Om Malik — Tue, 12 Aug 2008 21:11:31 +0000

Dr. Steven Low, a professor at the California Institute of Technology, has spent much of his life studying transmission ~~communication~~ control protocol (TCP), a technology that allows us to send emails, watch online videos and make Skype calls. TCP essentially makes the Internet (as we know it) work.

Through his work, part of a larger effort at the university, it was concluded that TCP, which came into existence in 1981, needed to be replaced by a more updated technology, one that could handle today’s Internet traffic. This updated version would take into account ballooning file sizes, the availability of fatter pipes, and the hyperglobalization of the Internet. In 2002, the researchers offered up this approach to TCP — called FastTCP — as a new standard.

That same technology is the bedrock upon which Low (as CEO) and Cheng Jin (as VP of engineering) have built FastSoft, a Pasadena, Calif.-based startup with $4.3 million in funding from Miramar Venture Partners and Caltech. (The company is currently meeting with VCs in the hopes of raising a new round of funding.) The two-year-old startup has developed a device that sits between a router and the Internet (or any other wide area network) and ensures the faster, smoother delivery of data.

FastSoft’s technology can be filed under the general “Internet acceleration” category along with the tech sold by vendors like Cisco, Juniper and Riverbed Technologies. But FastSoft’s true competitors are in fact content delivery network providers such as Limelight Networks. Typically a CDN has points of presence across the globe that are wired together for the specific purpose of quickly delivering information, mostly by offering up information from POPs that are closest to the Internet surfer.

By comparison, FastSoft’s technology allows you to get the same results by serving up information to one location from another, regardless of how far apart they are, without going through an expensive CDN. This is one of the reasons why tiny the startup counts the likes of Honda, Post Group, Reuters Australia and apparently Getty Images (though I haven’t yet been able to confirm this) among its customers. It also just snagged Technicolor, which, like Getty, is building what amounts to its own digital asset distribution network and will use FastSoft to deliver big production files to remote locations.

One of my sources tells me that FastSoft is also working closely with Internap, a company that operates data centers and a CDN. According to some of the experts I spoke with last week, a partnership here could be a game-changer for data centers, since buying and deploying FastTCP would save them a ton of money on bandwidth, a big factor these days due to the explosion of online video.

What’s Wrong With TCP? Nothing, But….

In order to understand FastSoft (and FastTCP), one needs to step back and understand TCP, specifically how it works and what its drawbacks are: In data transmission, a TCP client receives a packet; it then sends back an acknowledgment of that packet to the sender. Once the sender receives the acknowledgment another packet is sent, and so on. At any given time, an untold number of packets can be mid-transit.

This stream, Low explained to me, is where problems can crop up. TCP is based on the additive increase/multiplicative decrease (AIMD) algorithm, which slowly increases the speed if delivery is going smoothly, but when packets start playing truant, decreases the rate of transfer by half — sort of like going from fifth gear to second, without the option of gradually slowing down. (See the sawtooth-like pattern in the accompanying chart.)

All of which is fine when it comes to lightweight files such as emails or plain web pages, but makes the transferring of large files such as videos clips painful. As FastSoft executives pointed out, a 30 percent reduction in throughput on a 10 Mbps DSL line can mean the difference between an HD movie experience and a standard definition one. A 50 percent slowdown on a high-capacity line, such as an OC-3 connection, means the speed capacity goes from 155Mbps to a mere 77.5 Mbps.

What Makes FastTCP Better?

The idea behind FastTCP, Low explained, was to overcome the shortcomings of AIMD. He and his fellow researchers found that if they measured the time it took to send and receive acknowledgments from receivers, along with tracking lost packets, they could overcome the problem of traffic overflows and make the transfers smoother.

To that end, FastSoft has developed a box called the E-Series Accelerator, which figures out the fastest way to deliver packets to the end destination without flooding the networks, thus restricting packet loss to a manageable level.

The system can also use intelligence to figure out if the packet losses are random or if there’s a major underlying problem, like network congestion, and take appropriate preventive measures. Getting back to my automobile analogy, it’s like having a speed radar and a live traffic feed along with the ability to smoothly downshift gears.

Why It Matters

One of the reasons I got so excited about FastSoft is because there’s no technology required on the recipient site, nor does there needs to be any change to the current TCP paradigm.

To be sure, nothing is a slam dunk. There are a lot of things that need to go right in order for this company to make it. And they will undoubtedly be asked to explain the logic behind putting yet another potential point of failure inside the network.

Low, however, is confident that the cost savings (by not paying big bucks to CDNs) on infrastructure and bandwidth, along with the ability to ensure the smoother (and faster) delivery of data, is going to convince large companies dealing with big digital files to try out FastSoft.

Related research and analysis from GigaOM Pro:
Subscriber content. Sign up for a free trial.