Top Five Causes For Disruption Of Internet Service


slow-internetEditor’s Note: We took a poll from our customer base consisting of thousands of NetEqualizer users. What follows are the top five most common causes  for disruption of Internet connectivity.

1) Congestion: Congestion is the most common cause for short Internet outages.  In general, a congestion outage is characterized by 10 seconds of uptime followed by approximately 30 seconds of chaos. During the chaotic episode, the circuit gridlocks to the point where you can’t load a Web page. Just when you think the problem has cleared, it comes back.

The cyclical nature of a congestion outage is due to the way browsers and humans retry on failed connections. During busy times usage surges and then backs off, but the relief is temporary. Congestion-related outages are especially acute at public libraries, hotels, residence halls and educational institutions. Congestion is also very common on wireless networks. (Have you ever tried to send a text message from a crowded stadium? It’s usually impossible.)

Fortunately for network administrators, this is one cause of disruption that can be managed and prevented (as you’ll see below, others aren’t that easy to control). So what’s the solution? The best option for preventing congestion is to use some form of bandwidth control. The next best option is to increase the size of your bandwidth link. However without some form of bandwidth control, bandwidth increases are often absorbed quickly and congestion returns. For more information on speeding up internet services using a bandwidth controller, check out this article.

2) Failed Link to Provider: If you have a business-critical Internet link, it’s a good idea to source service from multiple providers. Between construction work, thunderstorms, wind, and power problems, anything can happen to your link at almost any time. These types of outages are much more likely than internal equipment failures.

3) Service Provider Internet Speed Fluctuates: Not all DS3 lines are the same. We have seen many occasions where customers are just not getting their contracted rate 24/7 as promised.

4) Equipment Failure: Power surges are the most common cause for frying routers and switches. Therefore, make sure everything has surge and UPS protection. After power surges, the next most common failure is lockup from feature-overloaded equipment. Considering this, keep your configurations as simple as possible on your routers and firewalls or be ready to upgrade to equipment with faster newer processing power.

Related Article: Buying Guide for Surge and UPS Protection Devices

5) Operator Error: Duplicating IP addresses, plugging wires into the wrong jack, and setting bad firewall rules are the leading operator errors reported.

If you commonly encounter issues that aren’t discussed here, feel free to fill us in in the comments section. While these were the most common causes of disruptions for our customers, plenty of other problems can exist.

Ten Things to Consider When Choosing a Bandwidth Shaper


This article is intended as an objective guide for anyone trying to narrow down their options in the bandwidth controller market. Organizations today have a plethora of product options to choose from. To further complicate your choices, not only are there  specialized bandwidth controllers, you’ll also find that most Firewall and Router products today contain some form of  bandwidth shaping and QoS  features .

What follows is an  all-encompassing  list of questions that will help you to quickly organize your  priorities with regards to choosing a bandwidth shaper.

1) What is the Cost of Increasing your Bandwidth?

Although this question may be a bit obvious, it must be asked. We assume that anybody in the market for a bandwidth controller also has the option of increasing their bandwidth. The costs of purchasing  and operating a bandwidth controller should ultimately be compared with the cost of increasing bandwidth on your network.

2) How much Savings should you expect from your Bandwidth Controller?

A good bandwidth controller in many situations can increase your carrying capacity by up to 50 percent.  However, beware, some technologies designed to optimize your network can create labor overhead in maintenance hours. Labor costs with some solutions can far exceed the cost of adding bandwidth.

3) Can you out-run your Organization’s Appetite for Increased Bandwidth  with a One-Time Bandwidth Upgrade?

The answer is yes, it is possible to buy enough bandwidth such that all your users cannot possibly exhaust the supply.  The bad news is that this solution is usually cost-prohibitive.  Many organizations that come to us have previously doubled their bandwidth, sometimes more than once, only to be back to overwhelming congestion within  a few months after their upgrade.  The appetite for bandwidth is insatiable, and in our opinion, at some point a bandwidth control device becomes your only rational option. Outrunning your user base usually is only possible where  Internet infrastructure is subsidized by a government entity, hiding the true costs.  For example, a small University with 1000 students will likely not be able to consume a true 5 Gigabit pipe, but purchasing a pipe of that size would be out of reach for most US-based Universities.

4) How Valuable is Your Time? Are you a Candidate for a Freeware-type Solution?

What we have seen in the market place is that small shops with high technical expertise, or small ISPs on a budget, can often make use of a freeware do-it-yourself bandwidth control solution.  If you are cash-strapped, this may be a viable solution for you.  However, please go into this with your eyes open.  The general pitfalls and risks are as follows:

a) Staff can easily run up 80 or more hours trying to  save a few thousand dollars fiddling with an unsupported solution.  And this is only for the initial installation & set-up.  Over the useful life of the solution, this can continue at a high-level, due to the unsupported nature of these technologies.

b) Investors  do not like to invest in businesses with homegrown technology, for many reasons, including finding personnel to sustain the solution, upgrading and adding features, as well as overall risk of keeping it in working order, unless it gives them a very large competitive advantage. You can easily shoot yourself in the foot with prospective buyers by becoming too dependent on homegrown, freeware solutions, in order to save costs. When you rely on something homegrown, it generally means an employee or two holds the keys to the operational knowledge, hence potential buyers can become uncomfortable (you would be too!).

5) Are you Looking to Enforce Bandwidth Limits as part of a Rate Plan that you Resell to Clients?

For example , let’s say that you have a good-sized backbone of bandwidth at a reasonable cost per megabit, and you just want to enforce class of service speeds to sell your bandwidth in incremental revenue chunks.

If this is truely your only requirement, and not optimization to support high contention ratios, then you should be careful not to overspend on your solution. A basic NetEqualizer or Allot system may be all that you need. You can also most likely leverage the bandwidth control features bundled into your Router or Firewall.  The thing to be careful of if using your Router/Firewall is that these devices can become overwhelmed due to lack of horsepower.

6) Are you just Trying to Optimize the Bandwidth that you have, based on Well-Known Priorities?

Some context:

If you have a very static network load, with a finite well-defined set of  applications running through your enterprise, there are application shaping (Layer-7 shaping) products out there such as the Blue Coat PacketShaper,which uses deep packet inspection, that can be set up once to allocate different amounts bandwidth based on application.  If the PacketShaper is a bit too pricey, the Cymphonics product can also detect most common applications.

If  you are trying to optimize your bandwidth on a variable, wide-open plethora of applications, then you may find yourself with extremely high maintenance costs by using a Layer-7 application shaper. A generic behavior-based product such as the NetEqualizer will do the trick.

Update 2015

Note : We are seeing quite a bit of Encryption on  common applications. We strongly recommend avoiding layer 7 type devices for public Internet traffic as the accuracy is diminishing due to the fact that encrypted traffic is UN-classifieble , a heuristics based behavior based approach is advised

7) Make sure  what looks elegant on the cover does not have hidden costs by doing a little research on the Internet.

Yes this is an obvious one too, but lest you forget your due diligence!

Before purchasing any traffic shaping solution  you should try a simple internet search with well placed keywords to uncover objective opinions. Current testimonials supplied by the vendor are a good source of information, but only tell half the story. Current customers are always biased toward their decision sometimes in the face of ignoring a better solution.

If you are not familiar with this technology, nor have the in-house expertise to work with a traffic shaper, you may want to consider buying additional bandwidth as your solution.  In order to assess if this is a viable solution for you, we recommend you think about the following: How much bandwidth do you need ? What is the appropriate amount for your ISP or organization?  We actually dedicated a complete article to this question.

8) Are you a Windows Shop?  Do you expect a Microsoft-based solution due to your internal expertise?

With all respect to Microsoft and the strides they have made toward reliability in their server solutions, we believe that you should avoid a Windows-based product for any network routing or bandwidth control mission.

To be effective, a bandwidth control device must be placed such that all traffic is forced to pass through the device. For this reason, all manufacturers that we are aware of develop their network devices using a derivative of  Linux. Linux-based is based on Open Source, which means that an OEM can strip down the operating system to its simplest components.  The simpler operating system in your network device, the less that can go wrong.  However, with Windows the core OS source code is not available to third-party developers, hence an OEM may not always be able to track down serious bugs. This is not to say that bugs do not occur in Linux, they do, however the OEM can often get a patch out quickly.

For the Windows IT person trained on Windows, a well-designed networking device presents its interface via a standard web page.  Hence, a technician likely needs no specific Linux background.

9) Are you a CIO (or C level Executive) Looking to Automate and Reduce Costs ?

Bandwidth controllers can become a means to do cool things with a network.  Network Administrators can get caught up reading fancy reports, making daily changes, and interpreting results, which can become  extremely labor-intensive.  There is a price/benefit crossover point where a device can create more work (labor cost)  than bandwidth saved.  We have addressed this paradox in detail in a previous article.

10) Do you have  any Legal or Political Requirement to Maintain Logs or Show Detailed Reports to a Third-Party (i.e. management ,oversight committee, etc.)?

For example…

A government requirement to provide data wire taps dictated by CALEA?

Or a monthly report on employee Internet behavior?

Related article how to choose the right bandwidth management solution

Links to other bandwidth control products on the market.

Packet Shaper by Blue Coat

NetEqualizer ( my favorite)

Exinda

Riverbed

Exinda  Packet Shaper  and Riverbed tend to focus on the enterprise WAN optimization market.

Cymphonix

Cymphonix comes  from a background of detailed reporting.

Emerging Technologies

Very solid  product for bandwidth shaping.

Exinda

Exinda from Australia has really made a good run in the US market offering a good alternative to the incumbants.

Netlimiter

For those of you who are wed to Windows NetLimiter is your answer

Antamediabandwidth

Behind the Scenes on the latest Comcast Ruling on Net Neutrality


Yesterday the FCC ruled in favor of Comcast regarding their rights to manipulate consumer traffic . As usual, the news coverage was a bit oversimplified and generic. Below we present a breakdown of the players involved, and our educated opinion as to their motivations.

1) The Large Service Providers for Internet Service: Comcast, Time Warner, Quest

From the perspective of Large Service Providers, these companies all want to get a return on their investment, charging the most money the market will tolerate. They will also try to increase market share by consolidating provider choices in local markets. Since they are directly visible to the public, they will also be trying to serve the public’s interest at heart; for without popular support, they will get regulated into oblivion. Case in point, the original Comcast problems stemmed from angry consumers after learning their p2p downloads were being redirected and/or  blocked.

Any and all government regulation will be opposed at every turn, as it is generally not good for private business. In the face of a strong headwind, don’t be surprised if Large Service Providers might try to reach a compromise quickly to alleviate any uncertainty.  Uncertainty can be more costly than regulation.

To be fair, Large Service Providers are staffed top to bottom with honest, hard-working people but, their decision-making as an entity will ultimately be based on profit.  To be the most profitable they will want to prevent third-party Traditional Content Providers from flooding  their networks with videos.  That was the original reason why Comcast thwarted bittorrent traffic. All of the Large Service Providers are currently, or plotting  to be, content providers, and hence they have two motives to restrict unwanted traffic. Motive one, is to keep their capacities in line with their capabilities for all generic traffic. Motive two, would be to thwart other content providers, thus making their content more attractive. For example who’s movie service are you going to subscribe with?  A generic cloud provider such as Netflix whose movies run choppy or your local provider with better quality by design?

2) The Traditional Content Providers:  Google, YouTube, Netflix etc.

They have a vested interest in expanding their reach by providing expanded video content.  Google, with nowhere to go for new revenue in the search engine and advertising business, will be attempting  an end-run around Large Service Providers to take market share.   The only thing standing in their way is the shortcomings in the delivery mechanism. They have even gone so far as to build out an extensive, heavily subsidized, fiber test network of their own.  Much of the hubbub about Net Neutrality is  based on a market play to force Large Service Providers to shoulder the Traditional Content Providers’ delivery costs.  An analogy from the bird world would be the brown-headed cowbird, where the mother lays her eggs in another bird’s nest, and then lets her chicks be raised by an unknowing other species.  Without their own delivery mechanism direct-to-the-consumer, the Traditional Content Providers  must keep pounding at the FCC  for rulings in their favor.  Part of the strategy is to rile consumers against the Large Service Providers, with the Net Neutrality cry.

3) The FCC

The FCC is a government organization trying to take their existing powers, which were granted for airwaves, and extend them to the Internet. As with any regulatory body, things start out well-intentioned, protection of consumers etc., but then quickly they become self-absorbed with their mission.  The original reason for the FCC was that the public airways for television and radio have limited frequencies for broadcasts. You can’t make a bigger pipe than what frequencies will allow, and hence it made sense to have a regulatory body oversee this vital  resource. In  the early stages of commercial radio, there was a real issue of competing entities  broadcasting  over each other in an arms race for the most powerful signal.  Along those lines, the regulatory entity (FCC) has forever expanded their mission.  For example, the government deciding what words can be uttered on primetime is an extension of this power.

Now with Internet, the FCC’s goal will be to regulate whatever they can, slowly creating rules for the “good of the people”. Will these rules be for the better?  Most likely the net effect is no; left alone the Internet was fine, but agencies will be agencies.

4) The Administration and current Congress

The current Administration has touted their support of Net Neutrality, and perhaps have been so overburdened with the battle on health care and other pressing matters that there has not been any regulation passed.  In the face of the aftermath of the FCC getting slapped down in court to limit their current powers, I would not be surprised to see a round of legislation on this issue to regulate Large Service Providers in the near future.  The Administraton will be painted as consumer protection against big greedy companies that need to be reigned in, as we have seen with banks, insurance companies, etc…. I hope that we do not end up with an Internet Czar, but some regulation is inevitable, if nothing else for a revenue stream to tap into.

5) The Public

The Public will be the dupes in all of this, ignorant voting blocks lobbied by various scare tactics.   The big demographic difference on swaying this opinion will be much different from the health care lobby.  People concerned for and against Internet Regulation will be in income brackets that have a higher education and employment rate than the typical entitlement lobbies that support regulation.  It is certainly not going to be the AARP or a Union Lobbyist leading the charge to regulate the Internet; hence legislation may be a bit delayed.

6) Al Gore

Not sure if he has a dog in this fight; we just threw him in here for fun.

7) NetEqualizer

Honestly, bandwidth control will always be needed, as long as there is more demand for bandwidth than there is bandwidth available.  We will not be lobbying for or against Net Neutrality.

8) The Courts

This is an area where I am a bit weak in understanding how a Court will follow legal precedent.  However, it seems to me that almost any court can rule from the bench, by finding the precedent they want and ignoring others if they so choose?  Ultimately, Congress can pass new laws to regulate just about anything with impunity.  There is no constitutional protection regarding Internet access.  Most likely the FCC will be the agency carrying out enforcement once the laws are in place.

NetEqualizer Bandwidth Shaping Solution: Hotels & Resorts


In working with some of the world’s leading hotels and resorts, we’ve repeatedly heard the same issues and challenges facing network administrators. Here are just a few:

Download Hotels White Paper

  • We need to do more with less bandwidth.
  • We need a solution that’s low cost, low maintenance, and easy to set up.
  • We need to meet the expectations of our tech-savvy customers and prevent Internet congestion during times of peak usage.
  • We need a solution that can meet the demands of a constantly changing clientele. We need to offer tiered internet access for our hotel guests, and provide managed access for conference attendees.

In this article, we’ll talk about how the NetEqualizer has been used to solve these issues for many Hotels and Resorts around the world.

Download article (PDF) Hotels & Resorts White Paper

Read full article …

Comcast Suit: Was Blocking P2P Worth the Final Cost?


By Art Reisman
CTO of APconnections
Makers of the plug-and-play bandwidth control and traffic shaping appliance NetEqualizer

Art Reisman CTO www.netequalizer.com

Comcast recently settled a class action suit in the state of Pennsylvania regarding its practice of selectively blocking of P2P.  So far, the first case was settled for 16 million dollars with more cases on the docket yet to come. To recap, Comcast and other large ISPs invested in technology to thwart P2P, denied involvment when first accused, got spanked by the FCC,  and now Comcast is looking to settle various class action suits.

When Comcast’s practices were established, P2P usage was sky-rocketing with no end in sight and the need to block some of it was required in order to preserve reasonable speeds for all users. Given that there was no specific law or ruling on the book, it seemed like mucking with P2P to alleviate gridlock was a rational business decision. This decision made even more sense considering that DSL providers were stealing disgruntled customers. With this said, Comcast wasn’t alone in the practice — all of the larger providers were doing it, throttling P2P to some extent to ensure good response times for all of their customers.

Yet, with the lawsuits mounting, it appears on face value that things backfired a bit for Comcast. Or did they?

We can work out some very rough estimates as the final cost trade-off. Here goes:

I am going to guess that before this plays out completely, settlements will run close to $50 million or more. To put that in perspective, Comcast shows a 2008 profit of close to $3 billion. Therefore, $50 million is hardly a dent to their stock holders. But, in order to play this out, we must ask what the ramifications would have been to not blocking P2P back when all of this began and P2P was a more serious bandwidth threat (Today, while P2P has declined, YouTube and online video are now the primary bandwidth hogs).

We’ll start with the customer. The cost of getting a new customer is usually calculated at around 6 months of service or approximately $300. So, to make things simple, we’ll assume the net cost of a losing a customer is roughly $300. In addition, there are also the support costs related to congested networks that can easily run $300 per customer incident.

The other more subtle cost of P2P is that the methods used to deter P2P traffic were designed to keep traffic on the Comcast network. You see, ISPs pay for exchanging data when they hand off to other networks, and by limiting the amount of data exchanged, they can save money. I did some cursory research on the costs involved with exchanging data and did not come up with anything concrete, so I’ll assume a P2P customer can cost you $5 per month.

So, lets put the numbers together to get an idea of how much potential financial damage P2P was causing back in 2007 (again, I must qualify that these are based on estimates and not fact. Comments and corrections are welcome).

  • Comcast had approximately 15 million broadband customers in 2008.
  • If 1 in 100 were heavy P2P users, the exchange cost would be $7.5 million per month in exchange costs.
  • Net lost customers to a competitor might be 1 in 500 a month. That would run $9 million a month.
  • Support calls due to preventable congestion might run another 1 out of 500 customers or $9 million a month.

So, very conservatively for 2007 and 2008, incremental costs related to unmitigated P2P could have easily run a total of $600 million right off the bottom line.

Therefore, while these calculations are approximations, in retrospect it was likely financially well worth the risk for Comcast to mitigate the effects of unchecked P2P. Of course, the public relations costs are much harder to quantify.

APconnections Announces 50-Percent-Off Sale of New NetEqualizer-Lite


Beginning May 26, all customers purchasing a full size NetEqualizer 2000/3000 model will qualify for a 50-percent discount on the NetEqualizer-Lite. In addition, the offer will be extended to all existing NetEqualizer users who will also be entitled to the 50-percent discount on their first NetEqualizer-Lite purchase. This offer is valid until June 30, 2009. Limit two per customer.

As well as offering users the same services available through previously released NetEqualizer models, the NetEqualizer-Lite is Power-over-Ethernet (PoE), handling up to 10 megabits of traffic and 200 users. Furthermore, the NetEqualizer-Lite also serves to solve hidden node issues without customers having to change their existing access points.*

Although the core technology behind the NetEqualizer has not changed, with the latest release price point, many ISPs and businesses are deploying the NetEqualizer-Lite closer to end users, often directly behind congested access points.

After just over a month in the field, NetEqualizer-Lite users are reporting they can now easily increase Internet subscribers by 30 to 50 percent at once congested towers and AP sites. For example, a customer with an 802.11b radio now has 100 subscribers on his network and is still running smoothly. In the past, this customer’s norm for saturation stood at roughly 20 users, but he is now enjoying a 500-percent increase after installing the NetEqualizer-Lite. This is translating into both higher revenues and a more satisfied customer base.

The NetEqualizer-Lite lists at $1499. In addition to the 50-percent discount, we are also currently offering volume discounts. Pricing information on all other NetEqualizer models is available online at http://www.netequalizer.com. For more information, please contact APconnections at 1-800-918-2763 or admin@apconnections.net.

*Hidden nodes are a problem frequently encountered by commercial wireless operators that has previously been solved using APconnections’ AirEqualizer technology. The NetEqualizer-Lite’s capability to offer similar solutions is simply one of the multiple benefits of the technology for administrators of networks of many different types and sizes.

New Asymmetric Shaping Option Augments NetEqualizer-Lite


We currently have a new release in beta testing that allows for equalizing on an asymmetric link. As is the case with all of our equalizing products, this release will allow users to more efficiently utilize their bandwidth, thus optimizing network performance. This will be especially ideal for users of our recently released NetEqualizer-Lite.

Many wireless access points have a limit on the total amount of bandwidth they can transmit in both directions. This is because only one direction can be talking at a time. Unlike wired networks, where a 10-meg link typically means you can have 10 megs UP and 10 megs going the other direction simultaneously, in  a wireless network you can only have 10 megabits total at any one time.  So, if you had 7 megabits coming in, you could only have 3 megabits going out. These limits are a hard saturation point.

In the past, it was necessary to create separate settings for both the up and down stream. With the new NetEqualizer release, you can simply tell the NetEqualizer that you have an asymmetric 10-megabit link, and congestion control will automatically kick in for both streams,  alleviating bottlenecks more efficiently and keeping your network running smoothly.

For more information on APconnections’ equalizing technology, click here.

%d bloggers like this: