Showing posts with label Akamai. Show all posts
Showing posts with label Akamai. Show all posts

Monday, October 24, 2011

Mobile Broadband Faster than Fixed Line?

There is an interesting bit of data in the latest report from Akamai on global Internet usage. The global average fixed-line connection speed was 2.6 Mbps, and the global average peak connection speed was 11.4 Mbps.



Looking at mobile broadband connections, average connection speeds on known mobile providers ranged from 5.3 Mbps down to 209 kbps, while “average” peak connection speeds ranged from 23.4 Mbps down to 1.2 Mbps.



The interesting observation is that wireless broadband has the higher peak speeds, about double that of fixed line connections, with a variable “average” speed that in some cases also is twice as high as fixed-line connections, though such sessions are highly variable. When mobile broadband is slow, it is an order of magnitude slower than fixed line connections. Global broadband speeds

Is "4G Plus DirecTV" a Viable Alternative to FiOS?


Verizon Wireless seems to be cooking up an out of market “video plus broadband” plan, working with DirecTV. During its recent quarterly earnings report, Fran Shammo, Verizon Communications EVP said that the company was working on such an effort.

“You're going to see that come in the fourth quarter with the what we now call the Cantenna, which is not a commercial name obviously, but it's the antenna that we actually trialed with DIRECTV, which was extremely successful,” said Shammo.

Some will legitimately wonder whether that approach might even wind up being used in some Verizon markets where FiOS has not already started to be deployed. LTE plus DirecTV

There are some significant Verizon markets including cities like Boston, Buffalo, N.Y, Baltimore and Alexandria, Va. where FiOS construction has not started.

The obvious new question is the rational approach Verizon should take to upgrading its fixed-line network. There isn’t much doubt about optical access media being more resistant to some weather-related impairments than copper networks, nor is there much doubt that new optical facilities cost less to maintain than older copper networks.

But the business question is how much incremental investment ought to be made in the fixed network,  if video and broadband services can be provided using the wireless network. One might rationally argue that the cost of maintaining the fourth generation wireless network is lower than the cost of maintaining the FiOS network.

Obviously, if that is true then the avoided capital investment in new optical facilities is significant as well. That isn’t to argue that fixed and wireless networks are in any way equivalent in terms of absolute bandwidth. But there is a financial question.

If the expected revenue and operating cost advantage of FiOS, compared to 4G, does not provide the optimal financial return, then a wireless solution might be the most-rational way to invest new capital.

The problem is that voice is a negligible contributor to incremental revenue on a FiOS network, while video, though an important contributor of revenue, is not such a great contributor to profits. That leaves broadband, and revenue upside is tough.

That is not to say fiber to home facilities are unimportant, merely to say that they might not be the best use of capital for a provider that also is investing heavily in mobile broadband.

In fact, there is an interesting bit of data in the latest report from Akamai on global Internet usage. The global average fixed-line connection speed was 2.6 Mbps, and the global average peak connection speed was 11.4 Mbps.

Looking at mobile broadband connections, average connection speeds on known mobile providers ranged from 5.3 Mbps down to 209 kbps, while “average” peak connection speeds ranged from 23.4 Mbps down to 1.2 Mbps.

The interesting observation is that wireless broadband has the higher peak speeds, about double that of fixed line connections, with a variable “average” speed that in some cases also is twice as high as fixed-line connections, though such sessions are highly variable. When mobile broadband is slow, it is an order of magnitude slower than fixed line connections. Global broadband speeds

Tuesday, March 2, 2010

50% Faster Email Performance Using Akamai and AppRiver

The amount of time it takes users to connect and sync with Exchange was reduced by half when using an Akamai-enabled AppRiver Secure Hosted Exchange solution, reports Compuware Gomez.

In other words, not all email services are created equal. Some can use optimization techniques such as Akamai to improve performance.

"AppRiver's relationship with Akamai is an excellent example of how network optimization can create a differentiated and improved IT service, in this case hosted Exchange," says Peter Christy, Internet Research Group principal analyst.

The enhanced performance also is an example of why overly-strict "network neutrality" rules that do not allow any forms of bit prioritization are problematic. There are lots of reasons to allow users, application and service providers to differentiate services and features, and better performance is foremost among them.

The AppRiver service offers a "one-of-a-kind, optimized hosted Exchange" service. The AppRiver and Akamai service essentially "privatizes" the connection between the user and the Exchange servers in order to create a high-speed virtual private network for AppRiver customers.

"By implementing Akamai's IP Application Accelerator service, AppRiver securely provides fast and reliable hosted e-mail for both wireline and mobile users," said . "Akamai improves the overall global experience for mobile device users by optimizing the Internet and minimizing the impact introduced by oversubscribed wireless networks," says Willie Tejada, Akamai VP.

source

Sunday, December 23, 2007

at&t Says It Will Provide CDN Services


Earlier this year Level 3 caused a stir when it said it would enter the content delivery network (CDN) market with a radical pricing model: essentially offering the quality of service features at no incremental cost to what customers expect to pay for simple IP transit. And if you think about it, that's precisely what a CDN does: provide QoS features on top of dumb pipe. All of which should have, and did, raise fears about the health of the CDN market.

After all, if a contestant says it will give customers for free, what they today pay for, that's disruptive. Most recently, at&t itself said it was getting into the CDN business as well. Which should have caused another shudder: remember Northpoint, Rhythms Netconnections and Covad? They were the three independent providers in the nascent Digital Subscriber Line broadband access market. Of course, when the incumbent telcos decided broadband access was a business they had to "own," they simply moved to do that.

So are Level 3 and at&t a threat to market-leader Akamai? Right now it's hard to say much, beyond the obvious fact that competition is increasing in the space. One issue could ultimately be the size of the market opportunity, the reason for that being that a smallish market will favor specialists, while a large market will favor the larger telcos.

And it is not necessarily simply because scale economies might kick in. It is more a matter that large telcos tend not to do well in market segments that are small. Small markets never get the attention they might deserve in a large organization. So unless the market gets fairly sizable, a large telco simply will not invest enough to keep pace with smaller specialists.

So how big is the market today? As it turns out, that's a guesstimate of sorts.

Some things are hard to count. Unified communications software is an example. People who track these things like concrete measures: ports, servers, licenses sold. So how do you track "presence" features that simply are embedded in the basic functionality of an IP PBX?

Other markets aren't quite that hard to track, but still are fuzzy because mutltiple revenue categories get lumped together in the reporting. Streaming media services, as distinct from application acceleration, provides an example of that sort.

Dan Rayburn,StreamingMedia.com EVP, provides a reasonable way of current approaching the U.S. market size, though. Working backwards from benchmarks, Rayburn suggests the market for CDN services (but not P2P apps) currently is less than $800 million.

Internap's 2007 revenue is about $24 million. Limelight Networks generated about $105 million for 2007 and about $95 million of that was earned in the U.S. market.

Akamai probably generates $400 million to $450 million of its $625 million total revenue comes from their CDN services. Rayburn further guesses that U.S. CDN revenues amount to $300 million.

Level 3 wasn't in the market for much of the year, but might have earned $2 million or so.

VeriSign might have earned about $8 million for the year in the U.S. market.

Mirror Image, CacheLogic, Panther Express, CacheFly and Advection.NET taken together will do about $20 million in the U.S. market.

EdgeCast, CDNetworks and BitGravity combined did about $5 million for the year. Again, these are new services that didn't have a full year of operation to measure.

PEER 1, NaviSite and Ignite Technologies together collectively generated about $8 million.

All other smaller regional service providers providing small and medium sized businesses outsourced video delivery services sold under $20 million in 2007.

Thursday, October 4, 2007

Level 3 Attacks CDN Pricing

Level 3 Communications has been gearing up for a major assault on the content delivery networks business and appears ready to price such services at a rate that basically offers caching and downloading services for no more than the cost of buying IP transport. If, as expected, Level 3 prices content delivery at the same price as Ip transit, it could disrupt much of the market.

It isn't so much the disruption of profit margins: that already is happening as several dozen contestants now are slugging it out for some share of the growing market.

The bigger issue is how participants in the media, hosted applications and enterprise end user markets are able to change the way they do things if enhanced quality becomes an integral part of the IP transport they buy.

Level 3 hopes to have its streaming services ready by mid-November. At that point it will have a wider shot at disrupting the market for transport of real time services. Right now much of the market is focused on video content. But there are other real time applications and services that really would benefit from lower-priced and more capable delivery over networks that eliminate jitter and latency over the global wide area network.

Access networks on each end still are issues, but tail circuits also keep improving. As more applications move to "cloud-based" processing and storage, they will have to start having the "feel" of local desktop apps. CDNs will be part of that experience. And that's where the major impact will lie.

Tuesday, August 28, 2007

One Movie: You Blow Your Monthly Data Plan


Something's gotta give here: Akamai has rolled out a high definition television delivery service, capable of delivering a two-hour feature-length movie encoded at a bit rate of at least 6-8 Mbps, with a resultant file size of 5 Gbytes to 8 GB. For those of you with some popular wireless broadband accounts, that's pretty much your "acceptable use" level of monthly consumption of data! And that's just your problem.

Assume the same bit of content were delivered to enough households to create one Nielsen ratings point. That's 1,102,000 households. Which means the delivery networks would require 6.6 Terabits per second of sustained bandwidth, assuming zero latency and zero network congestion!

I don't care who you are, your pipes are getting to be too small. Local area network bandwidth at enterprises is growing smartly, but consumer bandwidth won't be far behind, if in fact consumer bandwidth does not soon eclipse enterprise bandwidth in at least the downstream direction.

One HDTV movie. Two hours. Your whole monthly acceptable use consumption. Obviously there's not enough bandwidth.

"Tokens" are the New "FLOPS," "MIPS" or "Gbps"

Modern computing has some virtually-universal reference metrics. For Gemini 1.5 and other large language models, tokens are a basic measure...