For our development server we don’t have any static content, and wanted to run Tomcat on port 80 for simplicity.

On Linux, authbind simplifies this process. Here are the steps required:

Step 1: Install authbind

Step 2: Make port 80 available to authbind (you need to be root)
touch /etc/authbind/byport/80
chmod 500 /etc/authbind/byport/80
chown glassfish /etc/authbind/byport/80

Step 3: Make IPv4 the default (authbind does not currently support IPv6). To do so, create the file TOMCAT/bin/setenv.sh with the following content:
CATALINA_OPTS="-Djava.net.preferIPv4Stack=true"

Step 4: Change startup.sh
exec authbind --deep "$PRGDIR"/"$EXECUTABLE" start "$@"
# OLD: exec "$PRGDIR"/"$EXECUTABLE" start "$@"

The structure of our data is a long byte array (spectrum measurements) with associated meta-data (sample parameters, geo-location, etc). Each spectrum sample is about 200,000 8-bit data points, and the average database record is almost 256 kByte. At 8 samples per second, this adds up fase. Because each sample represents a distinct measurement location and time, change-detection compression is not possible. However, much of the data is redundant (i.e. long stretches of vacant spectrum), and so each sample lends itself well to atomic compression.

Samples are persisted into a standard SQL database (MySQL or Derby depending upon platform).

To reduce storage requirements I applied GZIP compression to the raw data, while storing meta-data in discrete fields for index, query and retrieval.

Within the Java entity object, I inserted a GZIP shim that compresses and de-compresses the raw BYTE data whenever data is written to or read from the database.

Here’s an outline of the compress/de-compress technique, stripped of the JPA encapsulation:

public class testGzip {

public static void main(String[] args) throws IOException {
// Original Data
byte[] dataBytes = new byte[]{1, 2, 3, 4, 5, 6, 7, 8, 9, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 2, 3, 4, 5, 5, 6, 67, 7, 8, 9};
// A grow-able, memory-resident byte array
ByteArrayOutputStream byteArrayOutputStream = new ByteArrayOutputStream();
// A gzip processor to write into the memory array
GZIPOutputStream gzipos = new GZIPOutputStream(byteArrayOutputStream);
// Write the data to the gzip processor
gzipos.write(dataBytes);
// Close the ‘file’ – this adds the gzip checksum, etc.
gzipos.close();
// Debug – show what we’re compressing
System.out.println(“data bytes len ” + dataBytes.length);
int len = dataBytes.length;
System.out.print(” data: “);
for (byte b : dataBytes) {
System.out.print(b + ” “);
}
System.out.println(“”);
// Get the compressed data – this can be persisted
byte[] gzBytes = byteArrayOutputStream.toByteArray();
// Debug – show that the compressed data is different (and presumably shorter)
System.out.println(“compressed bytes len ” + gzBytes.length);
System.out.print(” data compressed: “);
for (byte b : gzBytes) {
System.out.print(b + ” “);
}
System.out.println(“”);
// A structure to read memory-resident byte arrays
ByteArrayInputStream byteArrayInputStream = new ByteArrayInputStream(gzBytes);
// A gzip expander
GZIPInputStream gzipis = new GZIPInputStream(byteArrayInputStream);
// A place to hold the uncompressed data — the length is carried over from the original data
byte[] unzip = new byte[len];
// Read the compressed data into the uncompressed byte array
gzipis.read(unzip, 0, len);
// Debug – show that the recovered data exactly matches the input data
System.out.println(“un-compressed bytes len ” + unzip.length);
System.out.print(” data un-compressed: “);
for (byte b : unzip) {
System.out.print(b + ” “);
}
System.out.println(“”);
System.out.println(“Input data matches Output data: ” + Arrays.equals(dataBytes, unzip));
System.out.println(“Compression ratio: ” + String.format(“%.2f”, (double) gzBytes.length / (double) dataBytes.length));
}
}

“In 1697, Isaac Newton received and solved Jean Bernoulli’s brachistochrone problem. The swiss mathematician Bernouilli had challenged his colleagues to solve it within six months. Newton not only solved the problem before going to bed that same night, but in doing so, invented a new branch of mathematics called the calculus of variations. He had resolved the issue of specifying the curve connecting two points displayed from each other laterally, along which a body, acted upon only by gravity, would fall in the shortest time. Newton, age 55, sent the solution to be published, at his request, anonymously. But the brilliant originality of the work betrayed his identity, for when Bernoulli saw the solution he commented, ‘We recognize the lion by his claw.’”

Snowe also said a spectrum inventory would be “the necessary first step” to creating a policy framework. She is a sponsor of S.649, the “Radio Spectrum Inventory Act.”

“I am optimistic that our legislation will pass the Senate early this year,” she said.

Article:

• Sen. Snowe Digs Up 100 MHz of Unused Spectrum

I have some experience here. While at Comcast I oversaw certification and evaluation of all network devices, including next-gen set top boxes like the DOCSIS STB, which uses IP as a control channel instead of the proprietary Moto & Scientific Atlanta back-channel protocols. IP provides the foundation for interactive middleware. Add a CPU and hard drive (DVR) and you have the makings of a real media center.

My company tried (and failed) to introduce a new middleware stack for just such cable boxes. After much effort and burned cash my company learned a few things the hard way:

• Cable companies are like the post office: they deliver someone else’s material and collect a fee for the service.
• Cable technology (especially STBs) are designed to protect content first, distribute it second, and make it convenient to watch third.
• Adding features to a set top box is like adding features to an airplane – the testing and certification requirements alone make it nearly impossible to inovate.

In summary:

If the FCC really wants to crack this nut, and foster innovation in media distribution and consumption, the platform is not relevant. Instead, they must focus on satisfying content protection requirements. Until they satisfy the content owners, innovation in distirbution and consumption is neither relevant nor possible.

Links >

• Cable Technology
  • OCUR has dead pants
• Future of the Set Top Box
• FCC: Comment Sought on Video Device Innovation
• ARS: FCC admits CableCARD a failure

The system is very simple: enable a set of SIP IP phones to make and receive calls through our own IP PBX. This was done through a PRI, which we replaced with a VOIP trunk (running SIP) with an external phone company (Broadvoice).

Here’s what we did (after some trial and error).

Set up and configure the Ubuntu 9.04 server with Asterisk:

Very easy: Just install Asterisk.

% sudo apt-get install asterisk

Make a back up of config files. In /etc/asterisk:

% sudo mkdir -p Defaults
% sudo cp * Defaults/

Configure Cisco 7960 IP Phones and Asterisk

There are two ways to configure the Cisco IP phones: via TFTP or from the key pad. Either way, you’ll want to set up the following:

1) On Asterisk server, edit /etc/asterisk/sip.conf

Add the following at the very bottom of the file:

[username] ; this is the phone’s log in username (use something cute like phone1)
secret=password ; this is the phone’s log in password
username=KeyBridge ; this is the displayed outbound caller id
type=friend
user=phone
context=default
qualify=200
nat=no
host=dynamic
canreinvite=no

Note: Do NOT use your Broadvoice phone number for the Cisco username. It will confuse Asterisk. Instead, use something cute like ‘line1′ or ‘frontdesk’.

2) On the Cisco IP phone

First unlock the phone: hit [settings], then [9] Unlock Config and enter your phone password

Then configure your phone:

[4] SIP Configuration

[1] Line 1 settings

[1] Name should be set to the ‘username’ you chose in the asterisk configuration above
[2] Shortname is what this ‘line’ will be called on the local display. Try ‘Line 1′
[3] Authentication Name is the same as [1]
[4] Authentication Password is the ‘secret’ you specified in sip.conf
[5] Shortname is what this ‘line’ will be called on the local display. Try ‘Line 1′ (duplicate?)
[6] Proxy Address the IP address of your Asterisk server
Hit the [Save] smart key to save the new configuration.

Getting your Broadvoice SIP password

A word of caution: The Broadvoice portal password you set during account creation is NOT your SIP password. To get the SIP password, log in to the Broadvoice portal, then click [Account] on the tab and [Show Settings] under the My Devices block. Your SIP password follows the ‘auth_password[1]‘ text.

Configure Asterisk to Broadvoice SIP Trunk

On the Asterisk Ubuntu server, edit SIP.conf.

In the [general] section, uncomment the following lines:

disallow=all ; First disallow all codecs
allow=ulaw ; Allow codecs in order of preference
allow=ilbc ; see doc/rtp-packetization for framing options
allow=gsm ; Allow codecs in order of preference

Also in the [general] section, add a register command using the following format. (Note – the Broadvoice instructions are wong – they tell you use use some craziness like PHONENUMBER@sip.broadvoice.com as your username – follow this template).

register => PHONENUMBER:PASSWORD:PHONENUMBER@sip.broadvoice.com/PHONENUMBER

Obviously, substitute your own phonenumber and password.

At the bottom of the file (before of after your Cisco phone – it doesn’t matter) add a Broadvoice entry as follows:

[sip.broadvoice.com]
type=peer
user=phone
host=sip.broadvoice.com
fromdomain=sip.broadvoice.com
fromuser=USERNAME
secret=PASSWORD
username=USERNAME
insecure=very
authname=USERNAME
dtmfmode=inband
dtmf=inband

Edit extensions.conf

Now we’re done with sip.conf. Close it and let’s create the necessary dial plan. It’s easiest to start from scratch. Since we made a backup in ‘Defaults’, delete the file and create a new one.

% rm extensions.conf
% vi extensions.conf

Here is my complete file:

[general]
static=yes
writeprotect=no
clearglobalvars=no

[default]
; Broadvoice Domestic Dialplan
exten => _1NXXNXXXXXX, 1, dial(SIP/${EXTEN}@sip.broadvoice.com,30)
exten => _1NXXNXXXXXX, 2, congestion()
exten => _1NXXNXXXXXX, 102, busy()

; Broadvoice International Dialplan
exten=_011.,1,dial(SIP/${EXTEN}@sip.broadvoice.com,30)
exten=_011.,2,congestion() ; No answer, nothing
exten=_011.,102,busy() ; Busy

; Forward inbound calls to SIP Phone
exten => USERNAME, 1 , Dial(SIP/CISCOPHONE,20,rt)

Again, obviously substitute your Broadvoice phone number for ‘USERNAME’ and your Cisco phone username for ‘CISCOPHONE’

Start/restart Asterisk and you should be able to make and receive calls.

% sudo /etc/init.d/asterisk restart

A few quick confirmations just to see everything is working:

% sudo asterisk -vvvvr

> sip show peers
> sip show users
> show dialplan

Last Monday Edmund Thomas, representing Dell and Microsoft on this matter, fired off a note essentially saying “don’t count us in.” What’s odd is that neither Dell nor Microsoft have been involved in any Database development discussions but all of their earlier ex-parte comments were incorporated wherever possible.

When we discussed the matter with Microsoft directly, they hadn’t yet read the documents, so this is probably just a what it seems: a clarification that Dell & Microsoft haven’t officially weighed in yet.

It’s worth noting that the letter takes no issue with the substance of our work.

August 24: DELL, MICROSOFT DISPUTE ‘INDUSTRY CONSENSUS’ ON ‘WHITE SPACES’ DATABASE

Dell, Inc., and Microsoft, Inc., told the FCC yesterday that they are “puzzled” by a recent ex parte filing submitted by Key Bridge Global LLC that mentioned there is “a general industry consensus” concerning implementation of a TV “white spaces” database. Key Bridge Global wants to be selected as a database operator. “Although Dell and Microsoft appreciate Key Bridge’s efforts to further the discussion regarding white spaces databases, Dell and Microsoft are puzzled by Key Bridge’s claim that its filing reflects a ‘general industry consensus’ regarding database implementation and operation,” the companies said in an ex parte filing yesterday in Engineering and Technology docket 04-186. “To avoid confusion in the record, Dell and Microsoft wish to make clear that they have not authorized Key Bridge to make any presentations to the Commission on their behalf, including any presentations regarding a putative ‘industry consensus’ on white spaces databases Moreover, Dell and Microsoft note that much of Key Bridge’s 83-page submission discusses specific protocols and standards, which are outside the scope of the White Spaces Order and should remain so.”

August 14: KEY BRIDGE OUTLINES ‘INDUSTRY CONSENSUS’ ON OPERATIONS IN WHITE SPACES

Officials with Key Bridge Global LLC recently met with the FCC’s Office of Engineering & Technology to present their “understanding of a general industry consensus” on a range of issues related to unlicensed operation in television white spaces, outlining details related to a database architecture, security framework, and interference protection methods. In an ex parte filing in ET docket 04-186, Key Bridge, a Virginia company that is seeking FCC authorization to operate a TV white spaces database, said that such a database must do the following: protect incumbent licensees; publish unlicensed availability on a nondiscriminatory basis; maximize customer convenience and transparency; and execute FCC regulations as a neutral party. KeyBridge said it continues to develop its own implementation of the general designs outlined in its filings and that it plans to conduct a field trial later in the fall.

References:

• List of ex-parte filings in 04-186 docket (look for 08/17 and 08/24 filing dates)

In March, Senator John Kerry (D-MA) introduced the Radio Spectrum Inventory Act (S 649) in the Senate. Earlier this month, that bill passed the Senate Commerce, Science and Transportation Committee.

Last week, Representative Henry Waxman (CA-30) introduced a companion bill — HR 3125 — in the House of Representatives; the bill has been referred to the House Committee on Energy and Commerce.

The bills, if passed, would mandate an inventory of radio spectrum bands managed by the National Telecommunications and Information Administration (NTIA) and the Federal Communications Commission.

The Senate version calls for an inventory of frequencies between 300 MHz-3.5 GHz managed by the two agencies, while the House bill would mandate an inventory of 225 MHz-10 GHz.

S 649

Senate Bill 649 states that the NTIA and the FCC would be required to inventory the spectrum no later than 180 days after the bill becomes law; after the initial survey, follow-ups would be required every two years. Both agencies would need to prepare a report listing the licenses or government user assigned in the band, the total spectrum allocation, by band, of each licensee or government user (in percentage terms and in sum) and the number of intentional radiators and end-user intentional radiators that have been deployed in the band with each license or government user.

Additionally, if the information is applicable, the report would be required to show the type of intentional radiators operating in the band, the type of unlicensed intentional radiators authorized to operate in the band, contour maps that illustrate signal coverage and strength and the approximate geo-location of base stations or fixed transmitters. The report would then be sent to the Senate Committee on Commerce, Science and Transportation and to the House Committee on Energy and Commerce.

The bill also mandates that both agencies create a centralized portal or Web site that lists each agency’s band inventories. This information would then be made available to the public via an Internet-accessible Web site. Both agencies would also be required to make all necessary efforts to maintain and update the inventory information “in near real-time fashion and whenever there is a transfer or auction of licenses or change in allocation or assignment.”

“Our public airwaves belong to the American people, and we need to make certain we are putting them to good use in the best interests of those citizens,” Senator Kerry said when he introduced the bill in March. “Last year’s 700 MHz auction resulted in $20 billion for the treasury and will create greater opportunity and choice for consumers and businesses that need broadband service. We also took a great step forward when the FCC established a way for unlicensed devices to operate in white spaces. These two initiatives are evidence of how valuable spectrum is and how it serves as fertile grounds for innovation. We need to make sure we’re making as much of it available to innovators and consumers as possible.”

HR 3125

Like S 649, HR 3125 calls for the NTIA and the FCC to issue a report on the inventory of spectrum no later than 180 days after the bill becomes law; after the initial survey, follow-ups would be required every two years. The House bill goes a bit further than S 634, however, calling for the two agencies to work with the Office of Science and Technology Policy (OSTP); this office advises the President on the effects of science and technology on domestic and international affairs.

The agency reports called for in HR 3125 would include the same information called for in the Senate version. Like the Senate bill, the House bill calls for the reports to be made available on the Internet and update the reports as needed. Both bills include an exemption for licensees or users if they can demonstrate that disclosure would be harmful to national security.

“The [bill] represents a significant step in making available more spectrum for commercial and wireless services. The more efficient use of our nation’s airwaves will increase innovation for wireless products and services and improve the connectivity of the American people,” said bill co-sponsor Representative Rick Boucher (D-VA-9). “As more people use wireless devices and as advanced applications require higher data rates over time, additional spectrum will be needed to accommodate growth. Wireless technologies can also play a critical role in bringing broadband to more consumers, particularly in rural areas.”

S 649 is co-sponsored by Kay Bailey Hutchison (R-TX), Amy Klobuchar (D-MN), Bill Nelson (D-FL), Mark Pryor (D-AR), Olympia Snowe (R-ME), John Thune (R-SD), Mark Warner (D-VA) and Roger Wicker (R-MS).

HR 3125 already has 17 co-sponsors: Joe Barton (R-TX-6), Rick Boucher (D-VA-9), Steve Buyer (R-IN-4), Kathy Castor (D-FL-11), John Dingell (D-MI-15), Michael Doyle (D-PA-14), Anna Eshoo (D-CA-14), Bart Gordon (D-TN-6), Jay Inslee (D-WA-1), Edward Markey (D-MA-7), Doris Matsui (D-CA-5), Jerry McNerney (D-CA-11), Zachary Space (D-OH 18), Cliff Stearns (R-FL-6), Bart Stupak (D-MI-1), Fred Upton (R-MI-6) and Peter Welch (D-VT).

Links >>

• S 649 IS Radio Spectrum Inventory Act (Introduced in Senate)
• HR 3125 IH Radio Spectrum Inventory Act (Introduced in House)
• Office of Science and Technology Policy

Full text of the article:

Ofcom decides on white-space parameters
Location, location, location and a bit of sensing
By Bill Ray
Posted in Mobile, 2nd July 2009 08:02 GMT

UK regulator Ofcom has been considering what restrictions to place on white-space-exploiting cognitive radios, and has concluded that a location-based database is the only way to be sure.

The conclusion comes in a statement (http://www.ofcom.org.uk/consult/condocs/cognitive/statement/) following up on the consultation Ofcom launched back in February, and concludes that detection mechanisms might one day work, but for the foreseeable future any device intending to operate in white-space spectrum will need to check its location against an online database to make sure no one else is around before it starts transmitting.

Advocates of utilising white space – bits of radio spectrum used to broadcast TV in one part of the country which lie unused in other places – have been arguing that devices can detect and avoid transmissions, though tests have demonstrated the difficulty of that approach.

The easiest approach is to have an online database of available frequencies and require devices to check before they start transmitting. That means white-space-using devices either have to be fitted with GPS, or be set up by an engineer – with the latter option being most likely as the restrictions mean that white space is only really going to be used for fixed point-to-point connections.

Assuming it gets used at all, that is: Ofcom’s statement makes no reference to the calculations of Adrian Payne, who argued last December (http://www.theregister.co.uk/2008/12/03/white_space_uk/) that even if transmissions are allowed in white space the density of TVs in the UK makes any commercial deployment impractical.

Ofcom’s almost religious dedication to the capabilities of cognitive radio makes such calculations irrelevant, as well as making it impossible to reject detect and avoid entirely, as the FCC has done in the US. So in the UK, it will be possible to deploy a cognitive radio without checking the database – though at very low power levels and only if you continue checking for other users at least once a second.

Ofcom also isn’t saying what criteria it’s going to use to judge usable detect and avoid devices, noting that “detection-only devices [are] likely many years away and hence there is little advantage in rapidly making the necessary regulations to licence-exempt such devices”. The regulator is, however, planning to publish proposals for location-based white space systems later this year, which should prove interesting. ®

Links >

• Original Article

–Alexander Hamilton, Report on Manufactures, December, 1791