Results 1 to 18 of 18

Thread: Inter Site Connection Quality Parameters for Moto GTR8000 P25 Systems

  1. #1
    Join Date
    Jan 26, 2013
    Location
    Somewhere extremely cold and white.
    Posts
    89
    Thanks
    61
    Thanked 50 Times in 26 Posts

    Default Inter Site Connection Quality Parameters for Moto GTR8000 P25 Systems

    Hello,

    Just wondering if anyone knows what kind of IP connectivity is generally specified when you're connecting an Expandable Site Subsystem to it's controller?

    You might have a few remote sites, each with a handful of GTR's, and to work together as a system, they need to be connected to the controller at some building somewhere else. So one has to think about what kind of connectivity to specify.

    We'll likely be buying some of this equipment soon, but I'd like to know "first hand" what to expect to have to provide in order to have decent service.

    Also, when connecting two completely separate P25 ISSI equipped systems (ie. one from my organization using Brand A and one from another organization using Brand B,) what kind of connectivity will be required?

    I realize that there will be manufacturer-recommended specs provided when we buy the systems, but sometimes you get some insight from accounts of first hand experiences. (ie. something like "we were told to have 384kbps for a 6 channel ESS to M2 core connection with no more than 20ms jitter and 25ms latency, but found things worked way better if we upped it to XXX.")

    Thanks in advance for any insights!
    --
    Jim
    "Prepare: The time to win the battle is before it starts." - Frederick W. Lewis


  2. #2
    Join Date
    May 23, 2012
    Posts
    785
    Thanks
    442
    Thanked 878 Times in 363 Posts
    Country: United States

    Default

    Generally the site links are "stepped down" to T1 circuits for transport, that way there is "guaranteed bandwidth". The "output" of the site router is a T1 circuit, which the customer can then translate back to ethernet for transport over IP. The reverse is done at the other end.

  3. #3
    Join Date
    Oct 08, 2012
    Posts
    280
    Thanks
    4
    Thanked 177 Times in 77 Posts

    Default

    You need to provide dedicated transport with minimal latency and no jitter. You can forget anything over the public Internet - you need T1, metro-E, wireless (with a very strong/resilient path), etc.

  4. #4
    Join Date
    Jan 26, 2013
    Location
    Somewhere extremely cold and white.
    Posts
    89
    Thanks
    61
    Thanked 50 Times in 26 Posts

    Default

    Thanks for the feedback folks, but a T1...

    Not all that it's cracked up to be these days, I'm finding. Often, you send T1 to the carrier, only to have that carrier convert it from "good old" circuit switched to packet switched stuff and it suffers the fate all IP traffic does when aggregated.

    Now, T1 over our own microwave links works fine. It's ours, all ours, and there is no competing traffic as long as the link is good. And our new MW IP/ethernet equipment is more of an ethernet bridge, and just as solid. It's when it disappears into the cloud that worries me. Sure, you ask the networking guys for high QoS, and pay for it through the nose, but I'm wondering if anyone knows any practical limitations or can share lessons on this. For instance, a high QoS IP link over 500 miles connecting a site to a controller and then back to the adjacent site so comms between some members of a talk group on site A and the other members on site B (having roamed away but still on the same system) have to IP over 1000 miles with the added time of processing at each of the three nodes...? Anyone done this and find issues? I'll probably be facing a situation like this soon, and like my mom always said "to be forewarned is to be fore-armed."

    Thanks!
    --
    Jim
    "Prepare: The time to win the battle is before it starts." - Frederick W. Lewis

  5. #5
    IL02 No Longer Registered

    Default

    What you're asking is a complicated set of questions. Engineering will design the system and provide you specifications for any non-leased/microwave site links. I think you will find that Motorola's specifications are very stringent.

    I'll try and offer some general thoughts here.

    First, if you're dealing with any 3rd party backhaul, especially using such for Ethernet site links, you're definitely going to be negotiating service level agreements (SLAs), and discussing things like committed information rate (CIR), allowable packet loss, and as you mention, jitter, latency and QoS. All this stuff should obviously be known and committed to prior to turning up the system.

    You ask about connecting an ESS to its controller. Let's assume for the moment that you're doing standalone sites (not simulcast). In that case, using an ESS, the site controllers are co-located in the first ESS rack (assuming you have a site where the number of channels exceeds a single rack). Did you mean, connecting the ESS to the Zone controllers at the master site? That would make more sense, and yes, as mentioned above, there will be specifications for any Ethernet site links between an RF site and the master site.

    If you did mean simulcast, then yes, your site controllers (and comparators, collectively known as the Prime site) can (and often will be) remoted from the RF sites, though it is usual to co-locate one of the simulcast Subsites with the simulcast Prime site. Keep in mind that high quality links within the simulcast subsystem are much more critical, and require tighter tolerances, than links between sites and the master.

    You mention having two RF sites in close proximity, both with a 500 mile backhaul to the master site. That seems really strange to me, but let's roll with it for a moment. The first thing to be aware of is that no matter how close two sites are to each other, and no matter how far away the master is, all control and audio packets get routed through the master site anyway. Even if you were using some form of mesh topology, no two adjacent sites would ever be talking directly to each other.

    Now, speaking from a strictly technological standpoint, and ignoring cost considerations, you really shouldn't have radio sites that are 500 miles away from the master site. If you have, I'm guessing, a master site on one side of a state, with a bunch of sites in close proximity, and a few sites all the hell and gone on the other side of the state, what you really need to consider is expanding the system into a multi-zone system and homing those far-off sites to a second zone/master site much closer to them. This means that the 500 mile backhaul now exists between the master sites, not the sites and the masters, and much improves local survivability of communications in case of trouble with that link.

    As far as ISSI, I don't know what other manufacturers spec, but here's a PDF on the Motorola website which offers some details:

    http://www.motorolasolutions.com/web...ata%20sht_.pdf

    Basically, you're looking at a server, a firewall, and then one router per ISSI connection.

  6. #6
    Join Date
    Jan 26, 2013
    Location
    Somewhere extremely cold and white.
    Posts
    89
    Thanks
    61
    Thanked 50 Times in 26 Posts

    Default

    Got answers today from an authoritative source. Looking at 600 Kbps from full ESS to core with high QoS (<20ms jitter, <100ms latency, packet loss?, from ESS thru core to ESS). Just posting here for posterity's sake.
    --
    Jim
    "Prepare: The time to win the battle is before it starts." - Frederick W. Lewis

  7. #7
    Join Date
    Jan 26, 2013
    Location
    Somewhere extremely cold and white.
    Posts
    89
    Thanks
    61
    Thanked 50 Times in 26 Posts

    Default

    Following up a little more...
    The connection between two master sites needs to start at just under 5 Mbps, and again, with at least some of that marked to be high QoS, as per above. With more than light traffic between them this will rise.
    --
    Jim
    "Prepare: The time to win the battle is before it starts." - Frederick W. Lewis

  8. #8
    Join Date
    Jan 26, 2013
    Location
    Somewhere extremely cold and white.
    Posts
    89
    Thanks
    61
    Thanked 50 Times in 26 Posts

    Default

    Ok, and now for a chuckle at myself... It's amazing how much you learn by being tossed into the middle of something and forced to swim or sink. I just re-read IL02's post above and was amazed at how much more sense it makes after grappling with this stuff for the past 3 months or so. All I can say is that I'm glad we didn't have to do the 500 mile backhaul thing. There was a time when that might have been a reality we'd have had to face, but a little luck came our way and we were able to avoid it. Whew...
    --
    Jim
    "Prepare: The time to win the battle is before it starts." - Frederick W. Lewis

  9. #9
    Join Date
    Jan 10, 2013
    Location
    eastern side of au
    Posts
    263
    Thanks
    117
    Thanked 186 Times in 102 Posts
    Country: Australia

    Default

    Interesting - so one has to assume the issue becomes when you have 2 dozen sites and around 12 freq per site! The maths will become scary.

  10. #10
    Join Date
    May 23, 2012
    Posts
    785
    Thanks
    442
    Thanked 878 Times in 363 Posts
    Country: United States

    Default

    Yeah - it's simple. Just buy another $1M master site, because 500 miles is just too far.

    REALLY?

    Sheesh...

    I think a little investigation will reveal that a vast majority of these sites using GTR series equipment in a 7.x environment are still using T-carrier links. We have 15 channel sites running on a single T-1 link, and that's 1.544MBPS... So if they are telling you 5MBPS, well, it must be a BIG system.

  11. #11
    Join Date
    Jan 10, 2013
    Location
    eastern side of au
    Posts
    263
    Thanks
    117
    Thanked 186 Times in 102 Posts
    Country: Australia

    Default

    Don't ever let a moto sales person and entourage get between you and the next purchase.

    We see it all the time - nothing beats capacity of links (except the price tag for providing it!)

  12. #12
    Join Date
    Jan 26, 2013
    Location
    Somewhere extremely cold and white.
    Posts
    89
    Thanks
    61
    Thanked 50 Times in 26 Posts

    Default

    Hi d119,
    T-1 links were stellar, when they were really T-1 links. Nowadays, though, when you buy a T-1 link, you're more likely to get ushered to an interface dressed up as a T-1 link to your equipment, but which is really an MPLS-based cloud behind the interface box. It comes out of your equipment as T-1, but once it gets past the demarc, it turns into God-knows-what, bouncing God-knows-where. Then, when your jitter and packet loss gets out of hand and the quality of the service you paid for isn't cutting the mustard, you're in for a fight with a new generation of technicians who call themselves engineers and sneer at the fact that you're old-school enough to think that T-1 is a decent data transport mechanism.
    Sorry for the rant. Been down a rough road by mixing the T-1 thing with the telco.
    But you're right. A 9600 bps voice channel, even with overhead, isn't some sort of behemoth. I'm thinking that the massive data throughput requirements are excessive. ToS - I don't have an issue with tagging traffic with higher ToS. But requiring more bandwidth than you need to watch a movie on your internet connection - now that's a bit crazy. And for the record, bigger bandwidth can sometimes slow things down if you've moved into huge frame territory. Reassembling a massive frame at the other end can have very counter intuitive effects. (Of course, we had to learn that the hard way...)
    But if you've got a real, honest circuit-switched T-1 rather than the proverbial hogs ear dressed up as a silk purse, then I'm thinking you may be in the minority - and the luckier for it!

    On another note, I got a real smile today when one of our techs got hold of me and told me how well the GTR's were working out. After a recent replacement of their old stuff with GTR's, their workable RF footprint has grown massively. He's so amazed that he's getting his users to give him reports after they come back from far-away trips, just to verify it. I'm half-scared that we'll have the spectrum police knocking on our door saying "what gives?" We're not running outside our authorized spec, but if we end up causing havoc over a hundred miles away, ... well, I won't be totally shocked when I hear the knock on the door.
    --
    Jim
    "Prepare: The time to win the battle is before it starts." - Frederick W. Lewis

  13. #13
    Join Date
    Jan 26, 2013
    Location
    Somewhere extremely cold and white.
    Posts
    89
    Thanks
    61
    Thanked 50 Times in 26 Posts

    Default

    Wish me luck. It's factory acceptance test day today! <rubbing my hands together like a kid looking into a candy shop> I'll see if I can't find out what the effect of throttling back the intersite and inter-core links has.
    --
    Jim
    "Prepare: The time to win the battle is before it starts." - Frederick W. Lewis

  14. #14
    Join Date
    Jan 26, 2013
    Location
    Somewhere extremely cold and white.
    Posts
    89
    Thanks
    61
    Thanked 50 Times in 26 Posts

    Default

    OK, wow! The factory testing was pretty intense. Moto's Schaumburg facility is quite amazing. We toured their innards, played with the equipment and took our new cores and RF sites for a decent spin. There's nothing like some "hands-on" to make the ideas in the manuals come alive. Plus, it was an excellent way to discover things which were not entirely obvious from the literature - and, frustratingly, not easily found on the web. For instance, the connectors required to bring battery power into the expandable site subsystem are ...? Answer: none. Just bare the cables coming from the battery (tin it with the massive horsec#@k soldering iron if you're a bit "old school" like me) and clamp down the cable-ends inside the top of the rack's power entry panel. This one had been bugging me for a while. But after a thorough inspection, it's not a mystery any more.

    Pictures weren't allowed, but they provided a photographer for the traditional "smiling customer" shots so I put him to work taking a pile of shots of the equipment we're buying. I'm a bit obsessive about pictures, especially after I sent a tech to the top of a mountain to do an inspection of a shelter and he came back with about 10 photos, each being low resolution - only about 220kb each - not enough to answer so many questions I'd wanted to look into and which occurred to me only after he got back (ie. if we put equipment against that wall, is there anything critical on it that cannot be moved? - No photo, no answer. Poor photo, big question mark...)

    Anyway, I know many of you have been there before, but if you haven't, and you get the chance, don't pass it up! It's well worth the time. And really, when would you ever get to see all the stuff working together in the same room again? Pretty cool.
    --
    Jim
    "Prepare: The time to win the battle is before it starts." - Frederick W. Lewis

  15. #15
    Join Date
    May 13, 2012
    Posts
    613
    Thanks
    1,193
    Thanked 387 Times in 213 Posts

    Default

    Staging is pretty cool! The first time I was there, due to a scheduling issue I was there a day early. Lucky for me our FT was there, so while waiting. I helping them on another staged system. Until my system was ready.

    They had us broke up into 2 seperate groups, first week Sys Tech's /Eng, week 2 dog and pony show.

    Didyou get to keep you big banner, MOTOROLA Welcomes You?
    Radio Referenced...Those who think they know it all are very annoying to those of use who do.

  16. #16
    Join Date
    Dec 12, 2011
    Location
    Avalon
    Posts
    1,229
    Thanks
    353
    Thanked 367 Times in 176 Posts
    Country: United States

    Default

    More important did you get the free lunch? Much better in staging than in the Galvin center where you go for training.
    "Don't worry about what I am, cause I'm a state agent so what you need to do is make sure your doing the right thing **** boy" -J. Dewitte

  17. #17
    Join Date
    May 13, 2012
    Posts
    613
    Thanks
    1,193
    Thanked 387 Times in 213 Posts

    Default

    Jimbo, do you have any word on what encryption they are going with?

    I ask because of a lesson I learned the hard way!

    Imagine a year into the new system and the KVL goes DEAD. It is then we look and ask each other does anyone know the keys. Because if we send this thing in for repair, our loaner unit will be blank.

    The problem.....Motorola had loaded the keys into our KVL during staging, and like a dumbass all excited about the new system I failed to realize that we needed to create our own keys before we loaded up hundreds of radio's before deployment.

    And I did try things like all zeros, 12345 etc. it appears they just randomly just entered the keys. And that is good thing, but sucked at the same time,LOL

    I dreaded it and knew what the answer would be but I was desperate so I called Motorola and explained the situation to see if they could be of any help. They said quote "ARE YOU SERIOUSLY ASKING ME TO HELP CRACK YOUR KEY, THIS IS A PRANK RIGHT" LOL it didn't go over very well. I think they called the FBI and Homeland Security on me too!LOL

    It has a hard lesson learned, to say the least!

    But all is well in encryption land now!
    Radio Referenced...Those who think they know it all are very annoying to those of use who do.

  18. #18
    Join Date
    Jan 26, 2013
    Location
    Somewhere extremely cold and white.
    Posts
    89
    Thanks
    61
    Thanked 50 Times in 26 Posts

    Default

    We haven't implemented centralized keying, Moet. If subscribers want encryption, then currently, we're telling them that they have to manage their own locally. I'm going to revisit this if pressure from user agencies mounts, but at this point, it's the user's issue. Also, we've got some clients who don't want anyone handling their keys in any way anyhow. Fine with me, I say.

    Just for posterity in case someone Googles this up, and as it has some small bearing on the original title of the thread, I want to mention that at a particular radio site, getting data connectivity out to the shelter was problematic. There was POTS 25-pair copper in place, but no Cat5 or fiber. Solved the problem using a pair of CL1314R CopperLink Ethernet extenders from a company called Patton. Distance was about 3 km, give or take.

    The nice thing about it was that a local company, Luxcom, listened when I asked them if they had something along these lines (they're a local outfit which makes good fibre/Ethernet media converters which we use from time to time) but they didn't have something that fit the bill. Instead, they recommended asking Patton, and they were great! I don't know anything about Patton's other products, but these things work and the service was tops. This experience really cast Luxcom in a good light in my eyes, too.

    BTW, the system is now deployed and things are looking really good. Preliminary coverage tests are coming back two-thumbs-up. A big improvement over what we had before, and users are eager to get on board.
    --
    Jim
    "Prepare: The time to win the battle is before it starts." - Frederick W. Lewis