On Rad's Radar?

According to Om's article about P2P and Broadband Usage, Arbor Networks data shows that Pareto's Principle works:

On fixed and mobile broadband networks where consumer services are provided (i.e., NOT interprovider or typical dedicated Internet access for commercial enterprises):

• 10% of subscribers consume 80% of bandwidth.
• 0.5% of subscribers consume about 40% of total bandwidth.
• 80% of subscribers use less than 10% of bandwidth

How to deal with the 20% is the problem ISP's face. Comcast throttles P2P, which is about 20% of the traffic. Video is another big chunk. Caching can help a little. And no one wants FCC mandated network management rules. So what would be the answer?

The commentors across the web expect the bandwidth that is purchased. Apparently, none of them have designed, built or managed a network. Every network - even your home LAN - has a finite bandwidth. (Think about your own inside wiring in the house - for many it is so poor that it cannot carry the 100MB or the 1GB throughput that many homeowners want). Even HPNA taps out at around 50MB. There are bottlenecks in every network. Remember the busy signals when AOL stopped billing by the hour? The network wasn't built for that kind of usage. Just as the Internet is not really designed to carry real-time traffic such as voice and streaming video.

One management suggestion is to fire the 20%. No rule says you have to supply them with basically a dedicated line. (I would suggest adding that to any TOS and AUP you offer, just to spell out what exactly the service you provide is). Some ISP's have turned back to metering. This may be an answer, but it's like the cell phone owner who gets the crazy bill one month. In the old days, if you called international, you would get a huge charge from your LD carrier (at&t, mci or sprint) - and call them yelling. It would happen when the grandkids visit or they try to download a movie that takes 5 tries and ends up eating up 20GB. There's no one answer, except that the FCC should mandate. Use the FTC to enforce Truth in Advertising.

Having worked on committees, I can tell you that email and listserv usage make for a really challenging way to run a project. Now that I am in Chicago managing an MPLS migration, I realize that Microsoft doesn't have decent tools for this either. Outlook and MS Project help to schedule stuff and create a time-line, but when you want to look at spreadsheets or other docs, it's still cumbersome.

I have used a Wiki, which wasn't perfect, but it has a version manager, so you know what changed and by whom.

Yahoo Groups is pretty good for the communications side of things, because it contains a listserv, an archive, a schedule, a file storage section. Yet, Y! is a pita to get people without Yahoo! emails signed up. And it still doesn't do anything for document version management. Collaboration is a tough thing.

Emily Chang's Hub points out new apps every day. (LifeHacker.org and SarahinTampa.com points them out too). Emily has pointed me to STIXY, which is an online collaboration tool. (Free in beta). The invite people and version management are good. The GUI is like setting up Blogger. I have not gotten too deep with Stixy yet. (You have to get the team to buy in to using something they are not familiar with, which come to find out is difficult).

Another app that I have not tried yet is HomeCourt. This app (according to its front page) answers: Who is responsible for what; who said what and when; and where are the files for this project. Not all of it, but the responsibility part is pretty important, not just for CYA, but to find out who is holding up the project. Time is money.

I notice 3 years ago when my brother got his PMP certification, that PM kind of bloomed. However, it is mainly about how to create documentation. As someone not used to that, it is a cumbersome process (and truth be told, I can not figure out why you would have to spend so many hours just to deliver more status reports.) I guess the gold is in the project detail docs. In my current case, explaining what MPLS is and why it was chosen. How will it be integrated into the current network architecture. Define Class of Service; VLAN's; policies; IP Addressing schema; diagram the network and the NOC. These are documents that many data centers do not have. I can see why they don't (too many man-hours to produce) and why they should (easier to "franchise" the business. IOW, it is knowledge that you want to preserve in your business. When Dan leaves (or dies), what happens to all the knowledge about your network, your gear, your customers???? Gone. How do you replace that? Hence, why documentation is an important piece. (It's still one big hairball to create, but becomes an asset to your business).

We have seen Cogent have peering disputes with Level3 and Telia. There are scuffles regularly, usually about the size of the peering point. Obviously, the "Tier 1" providers would rather sell transit that offer Peering. There's no money in Peering.

With the mergers of MCI+VZ and BST+AT&T+SBC in 2005, conditions were made pertaining to the status quo on all peering connections being maintained. However, it was a three year term agreement -- and unlike most telco contracts, this one doesn't auto-renew. Already AT&T is re-negging on its peering inter-connection with, not some small guy, but with Sprint. [ DSLReports].

First, let me make the note that by Tier 1 carriers, I mean, AT&T, MCI-UUNET-VZB, Sprint, Level3, PsiNet-Cogent, Qwest, Global Crossing and Verio. You might be able to make a case for AboveNet or Savvis (C&W USA), but basically I went with the carriers that Savvis, InterNAP, Mzima, and TWIX aggregate and resell.

Meanwhile, the price of IP is seemingly dropping, especially as it is being bought in bulk more and more. (By bulk I mean OC-12 or larger ports, where the price can be sub-$10 depending on the vendor and commitment). But I wonder how all this will shake out as Peering Inter-Connection Agreements are re-structured. Maybe Level3 will be fine because they have the largest Looking Glass -- L3's network sees more ASN's than any other. AT&T and MCI are next. If Ma Bell is will to go to battle with Sprint, what do other ISP's have to look forward to?

Certainly the economics of selling sub-$20 per MB bandwidth gets skewed when all bandwidth is transit now, instead of maybe 50% now. Peering points will become much more important. (As will CDN's). Limelight and Google already allow networks to connect with them at carrier hotels like Equinix and Telx. From the user experience, as long as the ISP's bandwidth touches 55% of the ASN's needed for its traffic, the experience should be fine.

The Exaflood will be interesting because current peering points would break down under that kind of Internet Usage anyway. (AT&T thinks that the international carrier community has to spend $155 Billion by 2010 in order to upgrade networks for the coming flood. I wonder if they were counting on the IPv6 forklift upgrade in that figure).

And that brings me to my final three points:

One, how will peering work when one stream is IPv6 and the other is still IPv4? I'm sure Cisco has hardware for this, but that will add latency to the peering point.

Two, what happens with backlash? The Internet is really an inter-connected collection of autonomous networks (LAN's and MAN's connected to a great big WAN). As with Cogent's outages, the fallout is a Butterfly Effect.

Finally, since Inter-Carrier Compensation has been on the agenda at the FCC for about 10 years, I cannot foresee the FCC being much help here. In fact, as I observe Martin's dalliance with Network Management (over the Comcast torrent issue), all I can do is cringe. The FCC nows has 9 lawyers to every Engineer. Do you really want Lawyers deciding issues such as Network Administration and Peering? I don't, but Peering is on my Radar.

Every year, sales gets harder and harder. Finding qualified salespeople (the closers and hunters) is challenging. Then try to find someone who can lead or manage the salespeople. Sales Leadership is an elusive thing. One key reason is that sales training is missing.

We don't teach sales in school. Many companies have cut training budgets. However, sales is a process. If you want success, you develop and follow a system. From lead generation to qualifying the prospect to proposal to contract, there are a series of questions, actions, and follow up involved. Keith Rosen covers this in his book, The Complete Idiot's Guide to Closing the Sale.

In his latest book, Coaching Salespeople into Sales Champions, Keith Rosen gives every business execute the playbook to coach the sales team to be champions. Great athletes have coaches. And sales professionals know it all anyway. (Just ask them!) So it is more about Coaching than it is about micro-managing, if the desired outcome is a productive sales force.

Buy it for less than twenty dollars today or tomorrow and receive a bunch of bonus material from sales champions like Zig Ziglar and Tom Hopkins. Let me know what you think.

More and more video is moving to the web. Besides YouTube and all the wannabe's, porn, TV shows, movie downloads, and more. I'm sure you have seen the flash movie ads on sites like Forbes.com. The average web page used to be under 100K, now it is over 1MB. Mainly because so many sites monetize with tons of multimedia ads. I don't want to get on a rant here, but these pages can't be seen using Firefox with extensions -- and certainly don't work well on Mobile, which means you are missing the demographic you are making ads for. Duh!

Anyway, back to uses for video. Adding video to websites is the next phase of "cool web". It was the flash movie intro. Now it will be the video speaking intro. Sitepal will make you an avatar to use to customize the look and feel of your introduction.

But GoYoDeo recently introduced a unique video application that enables people and businesses to personalize who they are and what they are about throughout the social web. People can create videos that are superimpose over their profiles, blogs or web sites and deliver short or long messages then fade away. It's just one more example of self-expression.

And yet video email has not taken off yet. It's puzzling. Even with MLM from Talk Fusion pushing video email, I still don't see it in my inbox. Sightspeed and Tokbox both offer video email and video conferencing for free. Some video catched on (like on MySpace and YouTube) and some doesn't (like in email).

At Voicecon, John at Aastra showed me the nifty new Asterisk based PBX that was just introduced. Aastra also offers Microsoft's Response Point PBX, the key system replacement for 5 to 50 employees. The MS RP system is about $2750 -- about the same cost as PBX systems from other vendors in the Small Biz space. If Aastra offered me a choice of Asterisk or Microsoft, it would be a no-brainer to go Asterisk.

My Response to MS RP is that my experience with MS O/S offerings has been no fun. After the latest XP Updates, my audio driver and USB ports stopped working. (My USB ports are STILL not working -- hints anyone?). My Sprint PPC-6700 is an MS based smartphone that locked up on me while surfing Yahoo! Sports via IE. Remember Win98, SE, ME???

To me, placing my phone system in the hands of bloated coders who are newbies to voice is scary. I can barely get by with limited email, but if I don't have dial-tone, I could lose business. (Heck, MS IPTV makes me think that at some point I will miss the end of the game because of an update or reboot.)

I wasn't the only one feeling this way either. Read the review at Entrepreneur mag.

I will say now that MS and Cisco are fighting it out in Tele-Presence and VoIP, the noise level will spill over into mainstream. That means more opportunities to sell Hosted PBX, IP Phones, Unified Messaging, and Telco 2.0

We live in the Digital Age. Digital music, photos, e-cards, downloaded movies, etc. The mementos of life are now just electrons. Not many cards, letters, film. Nothing tangible.

Here's the kicker: Most people don;t have back-up either.

If you buy 100 CD's and strip off your MP3's and your iPod breaks or gets lost, you still have the CD's. When your hard drive fails, will you have DVD back-ups of your wedding photos? Your 40th birthday party? Or will it all be lost when BIOS can't find the drive? If your cell phone gets lost or broken, you have lost your address book. And in some cases have no way to call people since we are not dialing numbers any more - but contact links in the digital address book.

Despite Katrina, 9/11, and dozens of other weather based disasters, people still don't back-up data. One reason is that back-up is neither automatic nor easy.

To back-up an 80GB hard drive (the first time) using an ADSL line at the standard 378k upstream will take hours. Even storing your 2GB Outlook PST file can take a while. Add in the expense along with the hassle and the it-won't-happen-to-me mentality and you can see why there are so many back-up providers. None have become the Vonage of back-up. You know, huge ad spend with annoying music to explain what and why to back-up. The NYT had an article about one provider who I have never heard of: SugarSync.

Back-up is Insurance against the very real possibility of data loss. Hard drives do fail. Cell phones do get lost, stolen, and broken.

Back up is peace of mind. At least use a flash drive or external hard drive.

The world of landlines for the phone companies is a diminishing return. Telcos hold on to every revenue stream until the very end. I mean, Bell Labs discovered ISDN and DSL in the mid-1960's, but it took years to make it to market. Why cannibalize the $1500 T1 business?

Today, the landline voice business is diminishing. (Every quarterly report marks the decline). The cellcos have all released $100 Unlimited plans. Landline replacement might creep up.

One reason ILECs bundle DSL is to stall the slide. Only recently have consumers been given a choice of Naked DSL. That choice was mostly at the behest of the FCC in the AT&T-BellSouth merger. (However, AT&T makes the consumer jump through hoops to get it.)

Last week, Verizon was accused by the MSO's of misusing CPNI information for retention purposes. Customer Propriety Name Info was a hot issue last year over customer records being sold online. Telcos were supposed to tighten up CPNI rules. They have --- for everyone but their own employees, who have used the customer databases as a marketing tool.

Verizon, Sprint, and Alltel have spun off their landline business. Alltel became Windstream, which later acquired a smaller ILEC named CTC in NC. Verizon was able to "sell" off its rural New England division to tiny FairPoint. Sprint spun off its wireline business and named it Embarq. The focus is on cellular.

To stave off some landline loss, Embarq is rolling out a new landline phone - eGo.

"The new eGo product is a cordless phone which connects via DSL and offers visual voice mail and news flashes, Dow Jones reports. It will also offer a local business directory enquiries service and an online phone book." [source]

Another ILEC, CenturyTel has made some moves lately - buying GulfTel; acquiring 700 MHz spectrum; and selling a couple of markets to Zayo.

The landline business is consolidating to shore up declining revenue. (Telcos always think scale is the answer). At least, now a few telcos are trying to be innovative - with wireless overlays and new handsets.

My friend is kind of a patent attorney. He just got back from BarCamp Orlando (350 strong!). When he and I discuss Web 2.0, the one thing lacking is IP protection. Sure, everyone uses distributed computing in clouds and clusters, but no one stops to think about Intellectual Property rights.

It is more than just the ownership of the idea for something like Flurl or Vidoop. IP means licensing and asset management. There is a legal opinion that you shouldn't patent unless you can afford to defend it. The opposing opinion is to patent your idea so you legally own it.

When you look at the thousands of VOIP related patents, you know which way that sector went. It just seems like when web app dev is being discussed, IP isn't.

The other pitfall to this is Who Owns the Data? Lots of apps are moving to SAAS (software-as-a-service), where the data is kept on a 3rd-party cluster. How does Privacy work in that distributed model? Who is responsible for a security breach? These are factors that need to be considered as we expect SAAS to go mainstream. No business which owns licenses for, say, accounting software or Office is going to move to a web application that is not patent pending and without a clear delineation of privacy, ownership, and archive. These are just some things to think about when starting a business or application development.

While I was at FOWA, I learned a little about OpenID. I interviewed Kevin Fox, from VIDOOP, an OpenID provider. Want to learn more about OpenID? Listen to the MP3 of the Interview. (It was not my best interview, but Kevin is very enthusiastic about Vidoop.) In under 10 minutes you will have a good overview of OpenID.

You really should vet the material that you put in federal filings. Vonage learned that when they explained their technology enough in their IPO paperwork to enlist a patent infringement case from, well, everyone. In the Sirius-XM merger, it seems that neither company was operating within the 1997 FCC rules that established their charter. (How very RBOC of them). (see story)

Now Big Media (in this case Viacom) is yelling because the F agencies decided to approve the merger. It appears that Sirius doesn't compete with XM. Duh?

Consumers pick one or go back to old radio. Not many consumers would buy two different satellite devices. That's while the rule was that every device (satellite receiver) was supposed to be able to work with both broadcasters. -- And they did not. [Actually, according to Y! news, they did develop an inter-operable radio but no manufacturer would make it. It wouldn't be subsidized so who would buy it?

There were other rules ignored, including "You can't merge." Pish posh. Rules are for little people.

One thing many companies complain about is the price of bandwidth. The fact is price varies - greatly. Telegeography has a sampling of how much prices fluctuate. Why do they fluctuate?

Older contracts do not have declining prices, so people who bought a 5 year deal are paying more than people shopping now.

Where is the bandwidth? In about 8 cities - NYC, LA, DC, MIA, Dallas, ATL, CHI, and San Jose - bandwidth is really inexpensive (inside carrier hotels) because it is very competitive. In these cities, just about every carrier has capacity and wants to sell it to you. As you move outside those 8, capacity, availability, lit buildings, and competition change -- so too does the pricing.

The loop, the transport, the delivery of the bandwidth is the expensive part. It needs to be factored in.

In some of the examples, customers might have bought transit from companies during a period when they were having a fire sale or just plain wanted to take business away from another carrier. Or one carrier knew that capacity was limited (or the other carrier had implementation issues), so they could charge more.

If you need it now (and you want it in Ethernet instead of TDM (OC-x)) be willing to pay extra.

Looking for bandwidth? Drop me an email let me know what you are seeing.

Telecomweb reports on a study into the lobbying effort behind FISA bills. I'm only peripherally following the FISA bills. (David Isenberg is all over it on his Blog). The main gist is that the telcos want retro-active immunity for wiretapping everyone and everything. (So does the White House). But that kind of goes against the grain. It passed the Senate but thankfully the House is holding strong. (It would have been nice if the 3 senators running for office had voted on the FISA bill.

Microsoft is now ready to jump into VoIP for the SMB space with VoIP trunking and the Response Point phone system. Microsft joins Cisco, Linksys, a multitude of Hosted PBX players, Mitel, ShoreTel, and the rest of the hardware clans in competing in the SMB space.

My question is what space do they mean? Medium business of 500 to 1000 users? Small business of 250-500 employees or 100 to 250 workers? Or the under 100 employee segment that itself is segmented into under 25, 25-50, 50-100, and under 10. That's a huge sector that is heavily divided. Not just in size either. Every segment thinks about voice differently; has different needs; and must be sold to differently. I'll be curious to see what the marketing plan is. (So far I have not seen a plan to get to the SB space. Just the Medium size).

I had drinks with a VP of telecom for an IT distributor and we ended up discussing how the Hosted VoIP players will sell to the small business. There's not enough incentive for the headache. It will require the companies to throw money and manpower at it. Something they have been reluctant to do.

This space has been termed (by Seth Godin I think) the Fortune 5 Million because there are at least 5M small businesses with payroll in the US.

If you are playing in this space, drop me a note at peter at rad-info dot net.