Add a Comment Now - We Want to Hear From You
By Brian Boyko
Editor, Network Performance Daily
The Net Neutrality debate can often be overzealously couched by supporters and detractors of Net Neutrality legislation as an apocalyptic fight between good and evil.
Weirdly, they’re right. But not in the way you think.
See, at the core of Network Neutrality issues are appliances or programs which conduct traffic shaping. In traffic shaping, some packets are prioritized, others are held back. This prioritization can be done on the basis of content (what type of data is being transferred,) on the basis of application (what program is transferring the data) or on the basis of IP address (which computer is sending the packet, and which computer is receiving it.)
Now, here’s the rub: Traffic shaping can help improve network performance, decrease latency, and increase bandwidth by delaying those packets deemed to be of a low priority. Sounds good, right?
Not so fast. Traffic shaping can degrade network performance, increase latency, and decrease bandwidth… by the same means.
It all depends on who is controlling the traffic shaping and what packets they choose to prioritize. Traffic shaping, like guns, can be used for good purposes: A gun can be used for family protection, hunting dangerous or delicious animals, and keeping the King of England out of your face. And like guns, traffic shaping can also be used maliciously.
So, yes, Net Neutrality is very much a contest of “good vs. evil.” It’s just that the potential for good or evil depends entirely on who has control over determining what packets get priority, and for what purpose.
(Continued...)
In a corporate network, traffic shaping can be achieved by a number of technologies and techniques including Quality of Service policies and WAN optimization devices, and is an absolute boon. Of course the daily sales data should have a higher priority than the MP3 downloads coming from a sub-station of Sector 7-G. In fact, you couldn’t run a workable converged network with VoIP and critical data applications without traffic shaping because of the nature of UDP (used for VoIP calls) and TCP (used for data, and often for VoIP call setup) packets – without it, the UDP VoIP traffic could run amok on the network, crowding out TCP data packets until the entire network crawled to a standstill. Or conversely, without high enough priority, VoIP call quality could suffer because of packet loss or delay caused by bandwidth contention with data applications. Traffic shaping is one of the number one tools used by enterprise networks to get things working, keep them working, and most importantly, get them working most efficiently.
But corporate/private networks are not public Internet Service Providers. In a private corporate network, it is usually the corporation who determines the goals for the network and the IT department who figures the best way to use traffic shaping to implement those policies. An enterprise network may be worked on by many people, but ultimately, it answers to one master – the corporation’s bottom line.
With Internet Service Providers, it gets much hairier. An Internet service provider serves many masters – Your data is not as important to me as my data, and vice versa. We are all, as ISP customers, collectively selfish. That’s fine – we’re not maliciously selfish, we just only know what we ourselves are using our Internet connection for.
There are cases, though, even in an ISP, where sacrifices can be made by the individual for the greater good. VoIP and online gaming require low-latency connections; those applications and packets should be given higher priority over someone downloading large files – at least as far as latency is concerned. There are other examples as well and it is not hard to think of ways in which performance can be optimized to best serve all customers – someone’s video downloads in 20 minutes instead of 15 so that a hundred people wait only 30 seconds instead of 60 for checking their e-mail. Or maybe they wait 35 seconds so that 1000 people can send an instant message.
It is also not hard to think of ways this can be used maliciously – blocking applications which compete with the services provided by the ISP for additional charge, “blackmailing” online services into paying more to access the ISP’s customer base, even political censorship.
The current debate over network neutrality, the current apocalyptic rhetoric, and the clear “good vs. evil” battle is not centered on the conduct of ISPs but rather is centered on the question of whether traffic shaping should even be allowed.
To ban traffic shaping – or to adhere to a neutral network – is to assume that (all) ISPs, if given the chance, will use traffic shaping towards selfish ends – to disadvantage the customer in order to advantage the ISP. To support Network Neutrality legislation is to support the idea that this tool will only be used for evil and therefore must be eliminated. The potential and probable outcome is that the bad will outweigh the good of traffic shaping.
They may be right.
They are even probably right.
But to ban traffic shaping is to avoid the possibility of improvement – to pessimistically discard the idea that maybe the better angels of our nature will win out. Net neutrality is not only neutral in the sense of treating every packet equally – it is also neutral in the sense of being neutral in a fight between good versus evil.
And it is not hard to imagine different legislation that could preserve the good while expelling the evil. A competitive market would help, even though it’s not realistic. Broadband service is a natural monopoly, and where it is not, it is almost always an oligopoly anyway.
Perhaps traffic shaping could be harnessed for good if it was implemented by ISPs but not controlled by them – instead controlled by democratically elected, network savvy representatives of the user base. Or, more simply and probably more effectively, if the ISPs were required to base the upcoming month’s traffic policies based on the requested data traffic averaged over the past three months. These are just a few ideas and I’m sure that there are flaws in them – I’m not saying that they’re the be-all-and-end-all, and I invite others to create better solutions. The point is, legislation that enforces goodness, rather than neutrality, could be where we are focusing our efforts.
Network Neutrality may ultimately be necessary in the short term, in this limited political and economic landscape – but looking beyond the short term, it is possible to do much, much better.
For the sake of the human spirit, it has to be.
