Obama’s Speech: A Bandwidth Hog?

Barack Obama’s speech at the convention is available for download from the official convention site in what they term to be “high definition” video. And admittedly the video quality (making no value judgements about the rhetorical content) looks fantastic. But how much bandwidth is it taking up if someone were to watch Obama’s speech speech from last night – or John McCain’s speech next week.

We know that streaming video can cause extra traffic on the network – but exactly how much extra traffic? We downloaded “FreeMeter,” a utility with monitors bandwidth usage, turned off all our incoming and outgoing programs and launched the Democratic Convention Web site where you can watch the speech in high definition. The video, which is 54 minutes long, takes 1.11 gigabytes to download.

We’ll also look at the size of John McCain’s speech next week.

Which brings us to our next topic:

Wired: Comcast to make monthly Internet use cap official

Comcast has long been criticized for its unstated bandwidth caps – including by this blog. Getting hit by the FCC for the BitTorrent blocking might have affected the company, because for the first time Comcast has put a number on the cap:

On Oct. 1, the cable company will update its user agreement to say that users will be allowed 250 gigabytes of traffic per month, the company announced on its Web site.

Comcast has already reserved the right to cut off subscribers who use too much bandwidth each month, without specifying exactly what constitutes excessive use.

Bandwidth caps are poor ideas, as we mentioned previously, but at least this is more honest than cutting customers off who reach an unstated, unacknowledged limit.

Texas Tech Pulse: Interview: NetQOS Eyes IPO, Maybe

Texas Tech Pulse interviews the handsome, brilliant, witty, dashing, extraordinary, Steve “Danger” Harriman, VP of Marketing at NetQoS. [Full Disclosure: Steve Harriman is my boss’s boss.]

It’s regarding our press release of our earnings reports for Q1 and Q2 2008 revenues - typically not something that a private company put out, which fueled rumors that we were planning to go public. (To make it clear, however: NetQoS has made no announcements or filing to go public.)

Steve Harriman: We're now getting to a critical mass. If you want to do projections from the revenues that we announced this week, typically, Q3 and Q4 are even stronger than in the first half of the year. So our $60 million target for this year is certainly attainable.

As a company reaches that size, it's certainly no longer in the startup category. We're at that size where we have to consider future options. We have no current plans and nothing underway, but you start to think about the possibility of a public offering. Today's economic climate is not particularly conducive, but, obviously, a public offering is something that takes a lot of planning and preparation. We are very much in that phase.

Network performance is just as difficult – and just as important – as network security, but security is “sexier.” It brings to mind ideas of James Bond’s villain Boris yelling, “I am inwincible!” But, if you've got an IT staff that knows a lot about security but nothing about latency, you can guess how well the apps will perform.

But even separating network performance from network security isn’t enough – because the network fills so many different roles in the company, network engineers are becoming specialized by necessity.

According to IT Career Planet, Cisco just announced three new Cisco Certified Network Asscociate concentrations – CCNA Security, CCNA Voice (for VoIP issues), and CCNA Wireless – with an eye towards closing the “skills gap” and providing specialized knowledge. (Let’s side-step the whole “vendors offering training and certification” issue – that’s the way the system’s set up, and so far no one’s come along with a better solution to replace it.) Anyway, more specialized training is good news.

The bad news is that while Robert Whitely at Forrester Research says that in five years, organizations that have a dedicated position for wireless, voice, and security will grow as high as 70%, we can’t help but notice that he didn’t ask the question of whether there will be dedicated positions for network performance. Yes, it’s great that there’s going to be a VoIP specialization – but VoIP is only one of the applications that IT is delivering.

It’s one of the reasons that we’ve been offering (vendor-neutral) network training and certification in network performance technologies, metrics, and analysis, in our NetAnalyst program.

Scalability is one of the more overused terms in networking – which makes it hard to explain why it’s important. Well, I mean, beyond the main concept of: “More scalability means you can hook up more computers to it!”

True, how big the deployment is probably the best way to objectively prove scalability – for example, NetQoS has one ReporterAnalyzer deployment monitoring over 20,000 WAN links. No small feat. But scalability isn’t just the quantity of computers hooked up to the box, but also how much of the quality of the data you maintain when you’ve got tons of computers hooked up to the box. Or to put it another way, scalability means that in even large deployments, you get all the data at high granularity.

Talking about scalability in pure device count is sort of like talking about network performance purely in terms of fault. It is possible to have poor scalability without having no scalability, when you sacrifice detail for device count.

Another key of scalability that many people don’t think about is performance of the device itself. It would be ironic to purchase a device to monitor network performance that had a very slow UI because it strained under the load of monitoring thousands of links.

One of NetQoS’s many accomplishments over the past six months has been getting a patent on a memory management method and system which allows us to manage hundreds of thousands of combinations in a very small memory footprint.

Memory management is a major part of scalability, because allocating memory during a programming operation is relatively expensive, in terms of operating processor resources, to allocate memory during runtime. Put another way: the more efficiently you use memory, the harder you can push the processor on other tasks. For this reason, scalability requires efficient memory usage.

In addition to our own products, we also use it in our integrations into Cisco Wide Area Application Services (WAAS) – we’re able to integrate code there with little impact to the host systems.

Saying that your business should never, never, never use cloud-based applications instead of desktop or network/server based ones is about as ridiculous as saying that cloud-based applications will eventually replace IT completely.  

With an article that begins with “Cloud computing apps are for suckers. If there is an alternative that runs locally on your own machine, it will always be better,” John C Dvorak, seems to be going from “baiting Mac users” to “baiting Google users.”

But let’s just take the argument at face value.  Some of the points he makes are good ones – specifically, the ones with performance issues. 

I don't care if you have 30-megabit-per-second service—you'll get flaky performance from most online apps, especially if they're popular. Always remember that your online speed is only as good as the speed at which data is coming at you: The application server may be swamped, and the various nodes along the route could become clogged, too. Nothing is ever as fast as the machine sitting on top of (or beneath) your own desk.

Your desktop is faster than the cloud – that’s true - but is your car?  Information stored in the cloud can be accessed from any place with a Net connection.  Information stored locally can only be accessed locally – well, unless you connect through a VPN or set up a VNC server.  But even for those of us that know how to do it, a VNC server is a hassle, and a security risk unless you do it exactly right.  90 minutes is horrendous downtime for an enterprise application, and Dvorak is right so far as any application where 90 minutes downtime is unacceptable shouldn’t be put on the cloud. 

But there are plenty of applications – and for small-to-medium companies, e-mail is one of them – where the losses incurred from 90 minutes of downtime is less than the cost of having a dedicated in-house application installed and maintained on the network.  (If the opposite is true, don’t use cloud computing, use the in-house application, and keep an eye on how it performs.)

Dvorak also points out that your data is at the mercy of the service provider and that if the service is cut off, for whatever reason, so is your data.  That’s true, but if you don’t back-up your data, your data can be lost by a hard drive crash.  Both are about as likely to happen, in my experience. 

To Dvorak, “People tend to forget that software is NOT a service; the whole cloud scheme is a scam to lock users into a single product and somehow extract more money from them.”  There is some aspect of vendor lock-in, but mostly cloud computing is a way to provide an application at low startup costs in exchange for revenue over time – whether through advertising, in the case of Google’s apps, or through a subscription model.  Yes, it is very much “renting” rather than “owning,” but that can very well make financial sense in many cases. 

After that, the arguments get a bit silly. 

What happens if the net is attacked and your entire cloud world is gone for days and days? It just happened in the Republic of Georgia, and it can probably happen anywhere.

If the Russians start bombing us, John, I’m sure that the boss will give us a few days off. 

Ask yourself why the heck will we need six-core, high-performance chips if the cloud takes over everything?

Why do we need six-core, high-performance chips now?  In a virtualized server, certainly we’ll need power to spare, but unless you’re doing video editing or animation rendering, a six-core chip is probably overkill.  And if we stop putting the big iron in the datacenters of big companies (very unlikely,) they’ll pop up in the data centers of the SAAS providers. 

When it comes to performance and scalability, absolutely, standard client-server IT applications and local programs are going to have SAAS beat.  Final Cut Pro is not going to the cloud.  Photoshop isn’t going to the cloud (though Photoshop Elements is…).  But the key advantage of cloud computing isn’t performance or scalability – it is portability.  This is why people will pay twice as much for a laptop with the same specs as a desktop computer.  Mobility is important.   

I wanted to name this post “Things to view in Denver: Talking heads” but no one around the office got the reference.

Well, if you were worried that Obama’s VP announcement would overwhelm your network, you were spared. The announcement came in the middle of the night – 3 a.m., in fact – on a weekend. But convention season is starting, and that’s a whole additional set of worries.

At 5 p.m. EST today, the Democratic National Convention will begin, and with it, streaming video from multiple news networks and the DNC itself, which, like NBC’s Olympics coverage, uses Silverlight to project a “high-definition” (480p?) image.

Unlike the Olympics however, there’s bound to be some event – a protest that goes wrong, a verbal gaffe, a moving speech – that becomes a viral video. The reason I’m pretty sure about this is that it’s in the interest of both political parties that something goes viral – something good for the Democrats, something embarrassing for the Republicans. Politicians will find a compelling video of an unplanned, sincere, candid, spontaneous moment, even if they have to manufacture one.

There’s bound to be online coverage from the major TV networks as well.

And then after that, the Republican National Convention – with streaming through Ustream.tv - is next week. And McCain is yet to announce his vice presidential nominee.

Those interested in Obama and McCain’s technology policies will find this article at Ars Technica interesting, while C|Net has technology policy information for Biden.

Cisco has put up a new video in their “Seminar and Webcast Series” talking about “Energy Efficiency in the Data Center.” It may be produced by Cisco but the key points are pretty much vendor-neutral – starting with the idea that “Green” computing is a political/PR buzzword, and the way enterprises should look at the problem is one of efficiency and of sustainability.

Data center power consumption has more than doubled since 2001; the worry is that the trend will continue on an exponential pattern. This power consumption mainly comes from cooling the servers, rather than powering the servers; and with each 1U server (running 24/7/365) requiring the same amount of energy per year as it would take a Toyota Camry to drive 15,000 miles, energy efficiency is crucial.

Part of the solution is to buy more efficient components that cost more up front but pay money back. Another part of the solution is virtualizing servers, consolidating servers, and decommissioning servers.

They also mentioned using provided utilities to step-down the voltage if the server was underutilized – a trick laptop owners have been doing to get more life out of their batteries on the road. Same concept – if you don’t need all the power, consume less of it.

As far as the network goes, data center consolidation brought on by advances in WAN optimization is a big step towards reducing utility costs. Another step is taking advantage of the movement towards putting tools in the network infrastructure itself rather than as separate appliances – for example, putting SuperAgent network monitoring software (shameless plug) into Cisco’s WAAS.

These are all some common sense solutions and probably not the first time you’ve heard them. But the key point of the video-seminar was that just as we keep harping on the fact that you need to baseline your network performance to ensure that the changes you make to your network are having the desired effect, you also need to baseline your power costs as you make improvements.

I can’t believe I missed this the first time around.

I was so focused on how the online Olympic video was getting through the last mile, that I completely forgot to ask: How the heck are they getting it from Beijing to the U.S.?

Douglas Gourlay at Cisco has been blogging about how NBC’s been using Cisco’s Wide Area Application Services (WAAS) for WAN optimization, so that NBC’s video editors can use three 155Mbps OC-3 pipes, combined and load-balanced (with, of course, Cisco gear) to get the files directly from Beijing. While I’m not 100% sure on “as if they were stored locally,” holds true, it’s clear that WAAS is capable of some amazing stuff – we know because NetQoS has SuperAgent integration on WAAS devices and ACE load balancers. We track stuff like that all the time.

“This reduces operating costs of housing, air travel, transportation, and food. Avoiding 800 airplane trips also supports NBC’s green initiatives for the Olympic Games.”

It also probably makes the video editors a bit grumpy that they didn’t get to go to Beijing.

What I’m curious about is what will happen after the Olympics. Just as Olympic stadiums still stand – and are used – in every host city, I’m wondering if the infrastructure that NBC has to Beijing to deliver high definition video will remain after the Olympics. As China starts to become a new superpower, more news and information is bound to come from Beijing, after all.

And if this can be done for one series of events in one major city, is it that far off from having video-heavy WANs in every city to cover every major event?

And now, some additional notes.

Though Network Performance Daily is the company blog of NetQoS – it should be considered that this video and this posting are my own sole opinion as editor of the blog, and does not necessarily represent the opinions of NetQoS, their management, employees, or customers.

I know that many of our readers are D&D players; and most probably have a D&D player as a good friend. As mentioned in the video, we have had success explaining why those in IT tend to like playing D&D and using D&D as a metaphor for broader trends in IT. Since this blog owes D&D players some measure of its success, I felt compelled to speak up for those who play D&D.

As a general rule, whenever Network Performance Daily discusses politics, we do so from a viewpoint of looking at the issues involved that will affect network performance, or from a viewpoint of the IT structure of major modern campaigns, and not commenting on particular candidates or political issues. This case is a clear exception as I very specifically mention Sen. John McCain, the Republican nominee for President.

The reason we usually avoid commenting on political candidates and campaigns is additionally because this blog reflects on the entirety of NetQoS – and as you can imagine, among our employees and management, we have many McCain, Obama, and third party supporters. It would be inappropriate professionally – and insulting personally – if I was to profess to speak for all of them.

Nor do we think that Network Performance Daily’s readers would be interested in the political campaign season. NPD’s readers are typically interested in Network Performance, and unless it’s something like: “How to prepare for the traffic when Obama’s Veep is announced,” it just isn’t the appropriate venue.

But in this case, I do not think there would be any disagreement that Michael Goldfarb’s statements comparing liberal bloggers and New York Times editors to D&D players was meant to be a personal insult. We have no problem with Goldfarb insulting those he perceives to be against his campaign’s interests – that’s how politics is played, and liberal bloggers and the Grey Lady can defend themselves. What was offensive was using D&D players as the insult. It implies that playing D&D is somehow bad, wrong, or undesirable. That’s what prompted me to address this topic.

Goldfarb actually apologized on August 1^st for his conduct after the first incident when he called the New York Times editors. However, the fact that he repeated it on August 18^th shows the apology was less than sincere.

In the interests of fairness, we also did a search to see if Obama, or anyone on his campaign, ever posted, said, or implied anything negative about those who played D&D. We did not find any.

     -- Brian Boyko
     -- Editor, Network Performance Daily.

While we’ve talked quite a bit about what impact the Olympics may have on an enterprise network’s performance, we haven’t talked much about the performance of the NBC site hosting the live streaming of the Olympics. 

According to Jason Perlow at ZDNet, Limelight networks (which hosts the streaming videos) deployed the videos by going to the public internet by hosting the content more locally – at the ISP.  That means you’re viewing the Olympics through your ISP’s internal network, and the broader internet doesn’t even enter into the connection. 

This is smart thinking, it appears to be working, and by all measures this should be applauded.  Perhaps even duplicated – if you know that multiple employees will download the same content, local hosting on the LAN is preferable to duplicate download streams tying up the more expensive, slower WAN lines.

From the enterprise end of the equation, the fact that Limelight is delivering Olympics video more effectively just means that IT managers cannot count on their servers going down from being unable to handle the demand – IT managers still need to monitor their own networks for performance problems when a big event like the Olympics come up. 

However, it would be wrong to assume that Limelight’s strategy is the only reason why Olympic live-streaming hasn’t slowed to a trickle.

First of all, the site blocks 95.44% of visitors from accessing the content – because it limits the content only to those in the United States.  That’s a lot of people.

Secondly, the site requires Microsoft Silverlight. Most people don’t have Silverlight installed.  Some can’t even install it on their systems.  And there are certainly going to be a quite a few people who just didn’t think installing Silverlight was worth the bother to watch five minutes of Olympic footage they may be mildly interested in. 

And finally – none of the really popular sports are being streamed.  Gymnastics, Women’s Beach Volleyball, Swimming (with the exception of synchronized) and most of the track and field events aren’t available live. So you’re left with judo, fencing, and the decathlon.

So while it is a true technological wonder that the lights have stayed on and the site performs admirably – it is important to recognize that Limelight has not found a magic bullet to deal with extremely high internet video demand. 

By Ben Erwin

Product Manager, NetQoS.

After watching Jamaica destroy the competition in the male and female 100M races at the Olympics, I was reminded that I still have a crystal ball on loan from Miss Cleo. Needless to say, it was time to take another look into the future of network performance management.

After dusting off the ball, I began summoning the network performance spirits (via TCP/IP of course) and gazed deep into a phenomenon that really startled me. It seemed that IT managers were surrounded by performance management vendors claiming all of the capabilities necessary to manage application delivery.

But the spirits showed me that it was false!

Fancy user interfaces and marketing jargon were covering “checkbox” approaches to network monitoring. All of the vendors were touting NetFlow-based traffic analysis in a “me too” style of approach. Alas, none of them could scale beyond a few thousand interfaces and none of them could show much data beyond a few top 10 reports.

As the IT managers became frustrated with the lack of scalability and detail, the vendors came back to the table with a solution consisting of packet capture probes scattered throughout the data center and every remote site and really no answer to leverage NetFlow on a global scale. The sighs were deafening. When managers asked about integrating NetFlow with other forms of network performance data, piercing cries could be heard all over the world when vendors presented a lackluster version of their “portal” gris-gris with a separate tab for each data source – no reports actually integrated the data and IT managers begin growing increasing mad. But the bad mojo didn’t stop…

The boiling point was VoIP monitoring. VoIP had become the only option for electronic voice communication. Vendors were hopping from IT shop to shop showing off their new VoIP monitoring capabilities consisting of software agents. Instead of taking a passive approach to monitoring VoIP delivery from the data center, IT managers began sacrificing all of their free time and budget managing and deploying software agents all over the network. When IT managers asked the vendors about technologies like Cisco CBQoS or IP SLA for managing VoIP, vendors came back to the table with basic SNMP monitoring tools that did nothing more than provide basic MIB II statistics. Again, the vendors presented a checkbox, when a more robust solution was required.

I couldn’t stand it anymore as I knew IT anarchy was right around the corner. As more performance management vendors develop checkbox add-ons to their existing tools or acquire outdated technology such as agents and probes. I asked the spirits how IT managers will survive. Their answer was, “Reply Hazy, Try Again.”