Should we expect the same level of reliability in Internet service as we're used to in telephony and television? Surprisingly, we don't. We meekly accept frequent Internet service outages and faulty software. But as the line between telephony and Internet becomes thinner, more effort should be put into making Internet service as reliable as telephony used to be.
It's as if the ever-growing influence of PCs in our daily lives since the early 80s has conditioned us to allow for devices that fail or freeze for no obvious reasons. In fact, we seem almost happy that a simple flick of the on/off switch in most cases solves the problem.
In the old days, there was the “five nines” standard for telephone networks, meaning the service had to be available at least 99.999% of the time. This adds up to just over 5 minutes of downtime per year. Translated to individual user experience, if you make 10 phone calls a day, 99.999% reliability means it only happens once every 30 years that you pick up the phone and don't get a dial tone.
Data communications never had such a stringent requirement to start with, mostly because computers are not all that reliable themselves -- the term “nine fives” was sometimes used. Large companies that require high availability pay premium prices for service-level agreements (SLAs) or obtain multiple uplinks from different providers, but consumers have to put up with whatever our providers throw at us. But as data communications grew in volume and in importance over the past 15 years, the distinction between telecom and datacom vanished.
If Internet connectivity is to become as vital as telephony has been in the past century, meaning that outages have significant economical and social impact, now would be time for the industry to self-regulate and institute the requirement of five-nines reliability.
If you dissect the failure modes of our “Internet Experience,” there are three separate areas that require attention:
First, the terminals and customer premises equipment need to be more reliable. For example, our PCs, set-top boxes, cable modems, and (wireless) routers should have backup batteries to provide lifeline service, and the software (OS and applications) on all fixed and hand-held devices should be bug-free and stable.
Next, network infrastructure should be built with a high-availability objective. Individual switches, crossconnects, multiplexers, transmitters, and their software should have built-in redundancy, and the network should be designed to allow instant rerouting in case of equipment failure or cable cuts.
Last, but not least, the protocols on which the networks and services operate need to provide flawless access to all features and capabilities. In telephony Signaling System #7 is a monument of robustness that should set a standard for emerging frameworks like IP Multimedia Subsystem (IMS) and initiatives like TISPAN.
Some may say that networks have become more complex, hence more opportunity for failures. But I'd argue that providing a basic packet-based access pipe today is no more complex than it was 40 years ago to provide dial-tone telephone service. So if telephony could achieve five-nines reliability 40 years ago, we should expect the same for the Internet today.
First of all, it's my experience that Internet telephony (e.g. Skype) reliability is converging towards that of classic or wireless telephony, not so much because VoIP is improving, but because traditional telephony itself is degrading (or in the case of wireless, stagnating at an abysmally low level).
As for quality, there is no comparison - Skype blows landline or wireless telephony out of the water any day, as do other VoIP systems that support wideband codecs. For that matter, tin cans on a string blow wireless telephony away.
SS7 is not a monument of robustness. In fact, past a certain level of utilization, SS7 networks fail catastrophically, unlike IP networks that degrade gracefully. The only way to keep a SS7 network stable is to massively overprovision capacity. There have been a number of total telco network failures because they delayed expanding capacity in their expensive SS7 network, leading them to reach that critical threshold.
Phone networks do have a propensity for all or nothing failure modes. On 9/11, if you wanted to warn loved ones, the most reliable way was to use email, not overloaded cellular or landline networks.
There is no such thing as a free lunch - five nines reliability is very costly (even four nines), and is something you do not normally see until there is a failure. Thus reliability falls under the category of features that service providers cannot charge a price premium to recoup the investment costs for, because it is invisible. The inexorable effect of falling prices is
under-investment in reliability and falling standards, as in other markets like electricity. That affects both traditional and VoIP telephony.
Telephone reliability is a public good. When I was at the Paris Telecoms Institute, we were constantly reminded that when phones are down, lives are at risk. Unfortunately, the provision of such reliability cannot be left to the market alone, but regulation is not practicable either. The only way to bake in the incentives required would be to substantially raise the liability in case of downtime, and that would require more policy gumption than most governments are capable of.
As we become more and more dependent on the Internet for such basic things as phone service, the factors that lead to temporary failure of Internet devices and how to prevent it will become of greater and greater interest. It will become almost like a new branch of medicine, but having to do with machines rather than people.
A lot has to do with random behavior.
Sometime in early 2006 the word "random" became an "in" word to describe human behavior, "he's really random", means "unpredictable" but with many other shades of meaning too. Randomness can affect machines too, the handheld device that suddenly for no reason cannot connect to the Internet, the computer whose touchpad all at once becomes erratic.
The analogy between unpredictable behavior in people and machines is intriguing. Healthy normal humans regularly engage in unpredictable behavior, machines can also do so without being "broken". Restarting can sometimes return the normal state to computers, analogous to the cold shower in humans perhaps.
During a visit to Los Angeles in late August 2004 I had my Sony Aibo robotic dog in a carrying case in the back of a van parked on a hill. At a certain moment my host opened the rear gate and the carrying case accidentally went tumbling out with the dog landing hard on his head on the pavement. When I removed him from the case and pressed the on button he seemed very confused, tried to run all his programs at once, very much like the stunning effect of a human concussion. But then when I turned him off and then back on again he was fine. No permanent effects!
For several days in early 2006 the iPod nano in my car was unresponsive to its buttons and showed overlapping unreadable displays. A defective unit you say? Nonsense! A few days later it recovered and worked perfectly again. A period of illness, completely reversible, and part of "life experience" just like in people. One can postulate that a small ice crystal formed and then there was a small water droplet on the circuit board that evaporated, but you and I will never really know, just like a lot of human behavior is impossible to explain.
In Paris in August 2006 my Motorola V180 cell phone began refusing to charge. Everything else worked but it did not seem to recognize the charger, and would not go into charging mode no matter where or how I plugged it in or turned it on and off. This was a big problem, because as luck would have it my cell phone number was listed as the contact number on Leonard Cohen Night posters going up all over Edmonton just at that moment. The phone had to work! So I planned a desperate search all over Paris for a new charger that worked or worst case scenario purchase of a new phone I could transfer my sim card too so the same number would work in a new phone I could charge etc. I put together a list of possible phones to purchase and was just about to set out on this mad quest all over Paris when I accidentally knocked the cell phone on carpeted floor of the hotel room. I picked it up and plugged it into the charger and it made its happy charging noise started to charge normally! And no more misbehavior the rest of the trip! It was just being "random"!
As humans and machines merge (if they do!), the word "random" can increasingly be applied to both of them, and it is not entirely a bad thing. But we need insight into why temporary maladies of machines happen and how to prevent them!
I greatly enjoyed your piece on the randomness syndrom of humans and machines. The examples you cited brings reminded me of a similar one i had. Mylate fathet was a driver and i used to accompanied him as an apprentice during the weekends. On one fateful day we suffered a break-down during the last trip. We lost all the passengers on board as we tried fruitless toget the vehicle going so that we don't have to spend the night in the jungle. After hours of labouring on various parts of the engine, we finally gave up and had to settle to pass the night there. In the morning, to our utter amazement the engine was on at the first attempt without any effort of ours!! Back then we believe the breakdown wasn't a mechanical fault and when i read your piece, it just confirms my belief then.
My question now is do we have any scientific explanations for this randomness on the parts of machines? I know it is difficult to study or quantify random behaviour but at least some reasonanle explanations should be given on why such things do happen.
I know it will be too unfair at this point to expect internet reliability to that of telephony especially when the telephone has been with you for well over a century. But i wholeheartedly agree with you that with the manner in which we are shaping our cultures and systems around the internet, the issue of reliability should be tackle head-long. Reliability is critical to a variety of network applications like videoconferencing, streaming media and many others.
But reliability is only as strong as the weakest link in the chain and whilst it can be very high for one ISP, it can be significantly be affected when other ISPs enter the chain. So whilst it is easier to achieve higher levels of reliability in telephony, i don't see us attaining to the same reliability levels due to the complexities of the internet connectivity. What do you have to say about the efficiecy other methods being employed to increase reliability like SmartTunnel and overlay routing techniques??
What you said is indeed true? But the problem with Internet is that it is constantly evolving and the very nature of IMS and TISPAN is experimental to the best. Hence, it would take some more time for Internet to as reliable as Telephony system.
Diversification is the key, both for reliability and for safety and security. The price of devices continues to fall. If you are carrying one or more cell phones which can be used the moment the land line or VOIP phone fails, the absolute reliability of the land line or VOIP phone becomes less important.
Redundancy, diversification, and fail-proof security are needed, so if you lose one cell phone or other convergent device or it is stolen, you have another containing exactly the same stored information, and each device contain security features ensuring only you can use it. Multiple options, multiple devices like this is the ideal world, not one perfect device.
“Backing up” your life would consist of regularly copying the data in the cell phone or convergent device you carry to your backup device. That is the future we are heading toward, not one is which one device represents your one and only access to the phone system and your personal data.
I think the basic premise of telephony "reliability" is just wrong in modern times. Cell phones have changed that. Dropped calls are almost expected. No bars in not-so-urban places is annoying but hardly a grounds to change vendors.
The truth is, that occasional drop-outs or dropped packets is not the end of the world. The internet and telephony are not banking networks where dropped transactions are wholly unacceptable. The trade off of complete security or complete reliability is too high in terms of cost and redundancy.
My company sells internet appliances, and I often marvel at ISPs who spend a tremendous amount of money on redundancy, which just ends up complicating their network to the point where failures are actually more likely. I don't like it when my internet connection goes down for a few seconds, but I'm also not willing to pay double to avoid the problem. Total reliable is not necessary in either telephony or the internet.
We seem to be focusing on a middle level in these discussions. There are two other levels of interest: 1) the lower level "on the ground" questions of why sytems fail temporarily. In Paul Whyte's story one presuimes the engine was "flooded" (whatever that means!), in one of mine that there was frost on the circuit board, (but what explains the other instances?) and 2) the higher level "big picture" questions of what randomness really means and whether philosophically we must always expect Determinism in machines and Free Will in human beings.
My writing partner Nikki Olson writes:
"I think it is very interesting to compare random acts of machines to those of humans. For one, the Free Will-Determinism debate rests heavily on to what extent we find human actions 'random'. This becomes very important in political philosophy when we talk about things like rights and responsibility.To a great extent we find humans responsible for their seemingly random acts. We don't hold computers etc. responsible for their unpleasant random acts, we hold their human creators responsible. However, I will admit that in moments of anger and frustration with my computer my thoughts towards it can resemble thoughts I may have for a person. If I am angry with it for its random behavior, I may click the same button over and over so as to irritate it, give up on it and swear to buy something other than a Dell next time...Ultimately however I don't question whether its random action is truly random or consider it at all responsible for its random acts as I would another person. With sentient beings we may question something like 'intention', which we do not presently do with AI, but will perhaps be able to in the future."
"A significant question in philosophy, and the heart of the Free Will debate is whether there truly is something we can call random in this world. The word as applied to human behavior is ofter associated with acts of kindness, the mystery that is 'random acts of kindness'. Anyway..."
While I do agree that the internet cannot be 100% reliable, I do think we can expect more than what we get. Especially if we're basically being forced (good or bad) in to having to have heavy reliance on it.
From a consumer's point of view, I think better reliability can be expected if, for nothing else, what people are paying. Especially when you may or may not be able to get a refund credit if something goes wrong.
I guess, cost aside, reliability and what to expect would depend on what you use the internet for. If you're just a casual user, you might not think about it much if you get kicked off or you lose the connection for other reasons. But, if you rely on it for work, a lost connection is something that you cannot afford.
I do share your concern about cost but i think Leo's post raised a legitimate concern about internet reliability especially at a time when we are presently shaping our exixtence on the internet. With the increasing popularity of the Mobile web and other high bandwidth activities that rely on a reliable internet, i think your last statement may be a little bit off the mark. I can agree that end-end reliability may not be possible in either telephone or the internet but to say they are not necessary is to underestimate its significance.
So what will your advice be to those ISPs who will be facing massive reliablity challenges in the near future? Should they just take a passive stance or at least be seen doing something to alleviate the situation. I believe with better innovation, we may see a better reliable internet in the near future which is well position to meet our varying internet needs.
The ThinkerNet does not reflect the views of TechWeb. The ThinkerNet is an informal means of communication to members and visitors of the Internet Evolution site. Individual authors are chosen by Internet Evolution to blog. Neither Internet Evolution nor TechWeb assume responsibility for comments, claims, or opinions made by authors and ThinkerNet bloggers. They are no substitute for your own research and should not be relied upon for trading or any other purpose.
Using electronic stamps on email as a way to filter out spam or, at least cut down on unsolicited commercial email, has been considered by experts for some time now. Whether the discussion is raised in an email hoax or a serious proposal by a fellow ThinkerNetter, a practical solution is currently lacking. But what if the stamp charge is not collected by the government or ISP, but paid to the email recipient? The benefits there could be sufficient motivation to explore possible implementations.
People need social contacts. Right now we have to physically go somewhere to meet people socially. But suppose Internet-based social networking evolves to a point where everyone you know has a profile page with an integrated skype video account and a connection quality that enables a decent face-to-face conversation. Will that be a good enough substitute for water-cooler conversations (or beer talk, if you prefer)?
Internet traffic has been roughly doubling every year. As we know, this has created huge demand for network capacity. During the notorious telecom investment bubble, venture capital and R&D created a huge oversupply of network capacity. But was all that investment really lost? And what of the technologies created -- can we ever use them? I think we will soon learn that important optical networking technologies will be crucial to the evolution of the Internet. They just happened to have come a little bit early.
The Internet has opened up many new ways of doing business, and we've grown so accustomed to the convenience of electronic payments that we hardly miss the situations where we'd normally use cash: small transactions, tips, or the Salvation Army Red Kettle.
While Google introduces its new Chrome OS (which I'm hearing will be widely available in one year? Did I mishear that?), IBM announced 10 new products today to help companies using IBM System z mainframe technology.
Smarter Collaboration: How to Thrive in a Challenging Business Environment Market conditions are changing faster than ever, and organizations need to improve their agility and adaptability in order to provide better service and improve processes. The ability to work with customers, business partners, and employees as effectively as possible - while at the same time holding down costs - is a key to success. READ THIS eBOOK
your weekly update of news, analysis, and
opinion from Internet Evolution - FREE! REGISTER HERE
Wanted! Site Moderators Internet Evolution is looking for a handful of readers to help moderate the message boards on our site – as well as engaging in high-IQ conversation with the industry mavens on our thinkerNet blogosphere. The job comes with various perks, bags of kudos, and GIANT bragging rights. Interested?
To save this item to your list of favorite Internet Evolution content so you can find it later in your Profile page, click the "Save It" button next to the item.
How do you recognize an Internet bubble when you see one? Saunders explains how all bubbles have four symptoms in common – and takes a swipe at Google and Twitter into the bargain.
The sky is falling! And in other news, Saunders explains why he’s predicting a second Internet bubble – this one based around the current craze for social media.
China is investing heavily in fiber to the premises to propel itself into the world broadband Internet first division. What's it deploying, and what's it going to do with all that bandwidth?
Saunders predicts the decline and fall of America’s Internet empire, and explains how the Internet of the future will be multi-lingual as well as multi-national.
Saunders explains how Internet users in North America are already vastly outnumbered by those in the rest of the world – a situation which is only set to accelerate.
When Reiter gets incensed over incompetent Verizon FiOS order-taking and support, he broadcasts it via Twitter. Did it do any good? How should your company offer Twitter support? Watch this for all the answers.
The programmable Web, open APIs, and cloud-based services will fundamentally change orthodox telcos, and they need to decide what they are really good for in this new world: It's not necessarily what you might think, according to one of the world's biggest telcos.
Meatballs aren't an obvious source of monetization for the Internet, but Cloudy With a Chance of Meatballs may well be just that, and might also represent a new dynamic in the complex world of Internet content delivery.
The release of Microsoft's newest OS raises the question of the company's relevance in an era when Google dominates applications and search, and Apple runs circles around Redmond with its gadgets and user interfaces.
Industry initiatives and government stimulus funds are giving enterprise software vendors a great opportunity to help build out and manage smart grid technologies.
The problem with telepresence is that it's not universally accepted, because video calling isn't. While we can all do video calling, we also apparently worry too much about how we look. If we want HD telepresence in our future, we have to dress down, mess up our hair, and dive into our online life.
The US loses about $20 billion a year on pirated software, movies, and music. But public policy can help stem the tide of digital theft. For example, France has recently passed a 'three strikes and you’re out' law, whereby if after two warning letters an individual continues to download pirated software then his Internet access will be cut off. US policy makers should consider adopting similar policies.
Financial management planning does not need to include Voodoo economics, but it does help to tap into the knowledge base of your team through some sort of real-time system. We explore your options.
When Reiter gets incensed over incompetent Verizon FiOS order-taking and support, he broadcasts it via Twitter. Did it do any good? How should your company offer Twitter support? Watch this for all the answers.
The successor to the BlackBerry Bold 9000 – the Bold 9700 – will be available soon in the US. Is it worth upgrading? Reiter's got one, and offers advice.
Digg
Del.icio.us
Reddit
Email This
TWEET THIS
