Archive

Archive for February, 2012

Site to Site Streaming without breaking the bank – Test 1

February 22nd, 2012

About this time last year we were preparing to open our campus in Galesburg, IL.  Since the opening last spring this Campus has been using the recording of Saturday night’s teaching during their 11 am service on Sundays.  This solution has been a fairly stable, but has hasn’t operated without issues.  Additionally our pastor has really wanted to be able to teach the Galesburg campus live but we have multiple limitations… the distance between campuses is 50+ miles, our Peoria Campus is about 3-5 miles from any internet connections that can provide more than 10mb upload and any ISPs offering more speed has wanted nearly 6 figures in construction costs, and point to point connectivity is way more than we can afford. 

In addition to the current limitations the locations we are evaluating for future campuses don’t improve limitations on the list above.. in fact, they might be even more challenging.  Yet, being live is a huge desire from our leadership, so our quest continues.

Since Fiber isn’t an option at either of our campuses (but hopefully soon), we are limited to 50×10 cable modem in Peoria and a 16×2 cable modem in Galesburg.

our requirements for site to site streaming needs to:
-  provide 1080i display in the remote locations
-  not rely upon a private point to point connection
-  not require more than 10 mb upload from the sending location
-  be a solution that is easily reproduced for future sites in smaller towns with limited connectivity
-  be easily powered up and down by volunteers (not a 30 step process between multiple platforms).

We have demoed the Haivision Mako encoder/decoders and while the encoded video they produce is pretty amazing.. the pricetag is way to high to “not break the bank” not to mention  too high for for a “test environment” as we figure out what “live streaming” really means to our organization.  (However, you should at minimum demo their gear.. the Haivision gear gives a great benchmark for anything else you test.)

So we have been doing some testing with various other streaming solutions and thought we might share our mileage.

For our testing / phase 1 project we decided to try several pieces of gear:
-  Marshall VS-102 Encoder/Decoders
-  Wirecast and Wowza Streaming to a Roku
-  Marshall VS-102 and Zixi.com Hybrid

Our Media director thru the Church Technical Leaders and our peers at Willow Creek came across a encoder/decoder hardware (Marshall VS-102) made by the Display Monitor company Marshall Electronics (and Marshall USA).  We had heard that people were having good success using the VS-102 on the LAN but the device was capable of WAN streaming site to site.  The hardware is also able to additionally stream bi-directional Audio… (hmm maybe ClearCom in addition to the video’s audio?) This device across a LAN some pretty awesome results!  I came into the test expecting YouTube quality and was amazed.  If you are looking for a way to extend your HD/SDI video infrastructure this is a device you should checkout.  I don’t know of many hardware encoder/decoders in this price point … let alone something that can provide such a quality signal.

After a local LAN test, We quickly configured the boxes and streamed from site to site over our hardware VPN connections.  Remember we are using cable modems for our internet.. and the streaming at 1mb was solid.. but video quality was lacking… moving much above 2.5 mb we started to get a lot of jitter and audio drop outs.  If you have more than 10 mb upload.. I suspect you would have much better results, but those are just suspicions since we weren’t able to do such testing.

Next enter Chris Kehayias and his teaching us about Zixi.com.  Zixi is an internet based “private CDN” (Content Delivery Network), their strength is delivering HD video content over the public internet, including higher latency connections without the receiving end dropping frames or loosing quality or dropping audio.  The really awesome piece of the puzzle is the ability to stream from Zixi to a Netgear 550 Media Player.. (a endpoint and decoder for under $100 similar concept to roku).

So with all this new knowledge we started some field testing in Galesburg, so I thought I would share what we have tested and what our results were.

We first started streaming site to site with the two VS-102 units, with similar results to our pretesting, dropping frames and audio if we went above 3 mb.  Next we tested a roku streaming via the Amazon EC2 services but had stability issues even at 1 mb.  The quality of the video, when stable was pretty good, but couldn’t get it dialed in to keep a constant connection.  Next we configured the VS-102 encoder to stream a TS-Mpeg stream rather than the default streaming VS-102 to VS-102.  It streams to the Zixi “sending” application on a PC on the same network.  This PC is responsible for applying the Zixi goodness to the stream and sending it to their cloud.  Then at the remote campus we configured a Netgear 550 Media Player, pointed it to the stream and we have video.  We let the stream ‘chew’ for over 2 1/2 hrs, never dropped a frame or received the audio garbled. 

The only real test we couldn’t get working was to us the VS-102 decoder rather than the Netgear 550.  This was because we couldn’t get the VS-102 to receive what the Zixi receiver was pushing across the LAN.  We are working with Marshall support and expect to test this part soon.  Our motivation to getting the VS-102 to be the end point in Galesburg, 1 is the output of HD/SDI but also hopefully an improved video output beyond the Netgear Media Player.

After a fairly strong showing in Galesburg on Friday, we took the advice of Chris Kehayias, testing the stream during service to see the impact when our Wi-Fi is most populated and everything is buzzing… So during the Sunday Am services we tried streaming from the Peoria Campus to my house via Zixi, first 2 hrs total fail.. too much chewing thru our upload… and after smacking around a dropbox upload we were able to get a stable connection to Zixi and from there smooth sailing… even while the Sending Zixi software reporting having to recover over 50k dropped packets.  On the receiving end, you wouldn’t have known that Zixi was working so hard to keep the stream stable.

Overall I have been impressed by the flexibility of the VS-102, however their support has been limited.  The service of Zixi has been pretty amazing..  keeping a stream rock solid even with pretty poor ISP conditions.

Church IT, Hardware, Ministry, Tech , , , , , ,

Top Ten Reasons to attend the Church IT Roundtable 2012

February 20th, 2012

The 2012 top ten reasons to attend the Church IT RoundTable in Dallas Texas April 18th – 20th.

 

10. You have a project you need to finish by May 1st and don’t have a clue how to do it… nothing like 100+ free consultants to help you.

9. BBQ (nuff said)

8. North Texas spring thunderstorms are awesome!

7. LAN Party (Throw back to an old school LAN party) Bring your own DEW.

6. In & Out Burger AND Chick-Fila AND Hard 8 BBQ in the same city!

5. Free Workshops on Networking/Vmware, VOIP, Wifi and Exchange.  Yes I said FREE!

4. Trying out some of the cheapest most promising site to site streaming gear.

3. Building relationships with some of the most committed partners (aka vendors) serving the Church IT market who are ready to make your ministry shine.

2.  Because everyone else is going.. and you’ll be sad if you aren’t.

1. Hundreds possibly Thousands of Dollars worth of training and peer learning for only the cost of admission $75!

Don’t wait Register NOW!

Church IT, ChurchIT RoundTable, Ministry

Testing Lync Failover to Backup Registrar “Got Ya”

February 6th, 2012

Project Scope
Preparing for Deployment – Research and Education and Pricing
Deployment of Standard Server & Director Role
Deployment of Edge and Reverse Proxy
Deployment of Lync Voice Capabilities
Configuring Lync PSTN Calling thru Avaya IPOffice
Configure Lync 4 Digit Extension Dialing without DIDs
Configure Asterisk as a SIP Proxy for Avaya IPO and and Lync
Deployment of Lync Client to users
Testing Configuration of Backup Registrar
Training

Continuing the series in our Lync Deployment.  As we are approaching the date that we will completely cut over all users to lync we wanted to build in some redundancy to our deployment. 

We have done this by licensing a second standard server and configuring it in the topology as a backup registrar.  This will allow us to have a fail over server to host all voice calls in the event of a failure to the primary standard edition server (PSE).  The Backup Standard Edition Server (BSE) will provide voice capabilities and limited IM capabilities in a production down situation of the PSE. 
Note: for calls to be made in a ‘failed over’ scenario backup calling routes will need to be configured for the BSE mediation role as discussed in a future post

So we have configured our backup in the topology (how to in a future post) and configured the failover routes so it is time to test the scenarios.  For our testing we want to confirm that the PSE can fail and we can still make calls to the PSTN and if the PSTN is not available make a call out the analog backup lines.

You will want to review the default setting in your topology to set it to the lowest value possible when testing otherwise this test could take 15-20 minutes depending upon your value selected to fail over to a backup registrar. 

Failover

Our test was to remove the NIC from the PSE, the Lync clients will disconnect, attempt to re-connect and after the specified time connect to the BSE as the fail over registrar and make calls via the PRI and Pots lines.

However after configuring a Backup Registrar Lync Clients wouldn’t login during a failed server.  The clients would drop the connection as expected but however, they wouldn’t login to the backup registrar with limited functionality as expected. 

Side note… Kudos to @DHannifin helping figure this one out…
check out our awesome buddy Dustin’s blog:
http://www.technotesblog.com/ for lots of Uber good Lync goodness.

Even after changing the fail over time to just 30 seconds, the phone handset endpoints would login and calls could be made, but the Lync client would fail to login.   After some digging in the trace logs we found client that wouldn’t connect that we were getting an unauthorized error because the newly added BSE server wasn’t in the user certificate issued by the server to the client so the Lync client didn’t trust the backup registrar.

The Lync Client uses a certificate for communications with the front end server.  This certificate is not updated very often, in fact the default value to when it will update is 8760 HOURS that’s 365 DAYS!  (A little longer than we wanted to wait for our testing…Winking smile)

You can use the PowerShell command: Get-CSWebServiceConfiguration
to review the current values of your setting for MaxValidityPeriodHours’

CSWebserviceConfig

Since we didn’t have a year to wait, there are a couple solutions.
1. Change the default value by using the PowerShell command
Set-CSWebServiceConfiguration but this changes the cert settings for all clients and would require time for replication.
or
2. Delete the certificate on the machine that you are using for testing. This is a little more killing a fly with a sledge hammer, but for this testing appeared to be the best solution.

So in a testing scenario where you don’t want to change the re-issue certificate settings, on the machine you are using to test, simply launch an mmc window add the add-in for certificates and choose to manage users certificates.  Next browse to the personal certificates where you should find a certificate named the SIP URI of the user you are logged in as and it is issued by ‘Communications Server’. Delete the certificate and then restart your Lync Client (exit the application not just log off). 

Note: After deleting the cert, before you re-launch the Lync Client, you will need your primary front end server online so a new certificate can be issued to the client on the workstation.  Otherwise you still will not have valid certificate to connect and since the PSE is offline your client will try to connect to the BSE for which it still doesn’t have a valid cert.

After you re-connect to Lync to the PSE you can then power off the PSE (or remove the virtual nic from the virtual machine as we did.) You will notice the Lync client log off and after your Backup Registrar time out passes Lync will login to the Backup Registrar.  You will know this has happed when you see the Lync client display the red bar indicating limited functionality.

Lync Backup Registrar

If you have correctly configured a backup call route to your gateway, all voice calling will route out the gateway as if your Lync topology was operating normally.

Note: In an actual failover after you have configured all backup routes a call in progress should stay active even while the Lync Client is going thru its log off/log on process to connect to the backup registrar.  If you are in an active call during this fail over, your call should stay connected, BUT it will disconnect if you hit cancel on the Lync client during the reconnection process.

Church IT, Tech, Uncategorized , , , ,

National Church IT RoundTable 2012–Dallas Texas

February 1st, 2012

CITRT2012

I can’t believe that it’s that time of year again, but registration officially opens today for the National Church IT RoundTable (CITRT) April 18-20, 2012

The National CITRT  is one of those events that I plan to make every year… Thankfully 3.0’s release date is early March so I won’t feel TOOO guilty leaving my wife at home with a newborn and a 31/2 yr old.

The Church IT Round table gatherings are a great time to reconnect with old friends, meet some great new friends and learn from some of the great minds in Church IT.  I am already making my list of things I plan to learn while in Dallas.

Make plans now to join us at Watermark Church in Dallas.  This is going to be a can’t miss event!

JUST $75 includes most meals and the event registration for the 2 days (and optional training day).

This year we are adding a optional Pre-Event Training day that will include workshops on:
- Exchange
- VOIP
- VMWare & Networking
- and more

Also NEW this year, we’re making a focused effort to involve Church Web/ChMS developers and integrators, this is an important group of peers to our community, so this year we are building a track for web/dev.  We want our web/dev peers to benefit from an event like we have had over the past 6 years. The web/dev track will have breakout sessions to geek out about code, APIs, ChMS integrations, tools and best practices. These breakouts will be lead by web/dev heads David Drinnon (Second Baptist Houston) and Chris Kehayias (Calvary Chapel Melbourne). So go now and invite your web/dev peers on staff to join you this spring.

If you only go to one training/convention/workshop this year, CITRT2012 is the can’t miss event.. CITRT2012 will make you a better IT Pro, help you build relationships with others doing the same projects you are working on, and improve the ministry you support!

Check out the event info here: http://www.churchitnetwork.com/spring2012/

If you are a vendor with focus on serving the Church market and would like to partner with CITRT to help with this even or have space at the vendor expo check out the “Become A Partner” tab: http://www.churchitnetwork.com/spring2012/

ChurchIT RoundTable