Results 1 to 25 of 25
  1. #1
    Join Date
    Mar 2002
    Posts
    346

  2. #2
    Join Date
    Apr 2002
    Location
    Somewhere
    Posts
    17
    Ahh
    10Mbps Cogent Mulithomed Bandwidth
    care to explain that a little bit more, if it is only through one provider (Cogent) it is not multi-homed.
    Microsoft's new advertising slogan -
    "Buy a Microsoft Product, Help In The Downfall Of The Internet"

  3. #3
    Join Date
    Jul 2002
    Location
    Vancouver, BC
    Posts
    1,017
    yeah, i am confused... who are the providers for the multihoming?

  4. #4
    Join Date
    Mar 2002
    Posts
    346
    Epik is the secondary to the CogentCo bandwidth.

  5. #5
    Join Date
    Jul 2002
    Location
    Vancouver, BC
    Posts
    1,017
    but epik is essentially cogent bandwidth as well... so its cogent + cogent under a different name... heh

  6. #6
    Join Date
    Mar 2002
    Posts
    346
    netdude,

    You may want to recheck whereever/whoever you got that information from.
    It is incorrect.

    Epik resells Level 3 bandwidth, not Cogent.
    When you do a trace you will see this, Cogent isn't at any of the hops.

  7. #7
    srvoutsource

    your own server is hosted on epik, here a traceroute to your server:

    [...]
    6 p4-0.core02.dca01.atlas.cogentco.com
    7 p14-0.core01.atl01.atlas.cogentco.com
    8 p15-0.core01.jax01.atlas.cogentco.com
    9 p15-0.core01.mia01.atlas.cogentco.com
    10 g49.ba01.b002802-1.mia01.atlas.cogentco.com
    11 miamfllrh07-custbr-fastduck.ip.epik.net
    12 66.232.5.6
    [...]

    so SORRY but epik is ONLY the last hop before your gateway. If CogentCo network is down at hop 9 (as it IS often down in Florida), then your epik backup is useless.

    Right ?
    Last edited by neverless; 08-22-2002 at 05:22 AM.

  8. #8
    Originally posted by neverless
    your own server is hosted on epik, here a traceroute to your server:

    10 g49.ba01.b002802-1.mia01.atlas.cogentco.com
    11 miamfllrh07-custbr-fastduck.ip.epik.net
    12 66.232.5.6
    [...]

    so SORRY but epik is ONLY the last hop before your gateway. If CogentCo network is down at hop 9 (as it IS often down in Florida), then your epik backup is useless.

    Right ?
    Epik has peering agreements with Level3 and Aparently now Cogent! (That's brand new)

    By trying to prove Serveroutsource wrong and bash Cogent in the same sentence, you actually fail at both attempts. Your comment about Cogent being DOWN often in Florida is different from my actual experience over the past 6 months. Can you provide proof?

    What is actually being proven here and by those Tracert's is that Cogent has yet another quality Peering agreement. Therefore, strengthing it's Network. Therefore showing the value in Serveroutsource's posted offer.

    A complement for serveroutsource.net from a competitor, here on WHT???........ That's gotta be a first!

    If you would disclose the full tracert, everyone would see that before the cogent hops is a single psi hop to Level3 anyhow. Most all of the carriers are BGP connected via their peering arrangements, so your "OMG It's Down" doesn't hold much water anyway.

    I monitor my Network (Yes it's Miami Cogent/Epik too) from 3 different locations in the US (West Coast, Far NorthEast, and Midwest). What I find amusing, is that when one of my monitor's shows my Network as being Down, The others report no problem at all and my internal MRTG tracking shows the Network as UP.

    I can recommend Serveroutsource as being an Honest business person above the age of 18 (considered a minority here on WHT) with a quality offering in a Fine telco grade datacenter in Miami. Although I'd rather have the customer myself, the customer isn't going to get "taken" by using this particular competitor of mine.

    Go pick on the Kiddy's trying to scam people out of Hardware with their bogus colo offers. This guy's legit!

  9. #9
    Join Date
    Mar 2002
    Location
    Mass
    Posts
    726
    Your comment about Cogent being DOWN often in Florida is different from my actual experience over the past 6 months.
    I don't know about it being down, but I know it sucks.
    Jason

  10. #10
    Your comment about Cogent being DOWN often in Florida is different from my actual experience over the past 6 months. Can you provide proof?
    We had a 1/2 rack with WholesaleColo few months ago and had to cancel it due to connectivity problems. We had failures of 1/2 minutes everyday. This kind of failures may not be recorded if you check your connectivity each 30 minutes...

    BTW:
    http://www.Webhostingtalk.com/showth...threadid=68977

    If you would disclose the full tracert, everyone would see that before the cogent hops is a single psi hop to Level3 anyhow. Most all of the carriers are BGP connected via their peering arrangements, so your "OMG It's Down" doesn't hold much water anyway.
    A single hop ??? Kidding ? BTW look at hop 15... 449 ms

    1 dsl-64-34-232-6.telocity.com
    2 route-216-227-81-1.telocity.com
    3 216.227.80.124
    4 ge-1-2-0-50.core2.mia.tlct.net
    5 ae0-5.core1.mia.tlct.net
    6 so-1-0-1.core1.was.tlct.net
    7 ge-0-0-0.border1.was.tlct.net
    8 mae-east.cogentco.com
    9 * p15-0.core02.dca01.atlas.cogentco.com
    10 p14-0.core01.atl01.atlas.cogentco.com
    11 p15-0.core01.jax01.atlas.cogentco.com
    12 p15-0.core01.mia01.atlas.cogentco.com
    13 g50.ba01.b002802-1.mia01.atlas.cogentco.com
    14 miamfllrh07-custbr-fastduck.ip.epik.net
    15 66.232.32.2 (66.232.32.2) 128.225 ms 449.126 ms 225.849 ms
    16 66.232.32.6 (66.232.32.6)
    17 raq4less.com (66.232.63.20)

    I can recommend Serveroutsource as being an Honest business person above the age of 18 (considered a minority here on WHT) with a quality offering in a Fine telco grade datacenter in Miami.
    Hey, I don't have any contract with srvoutsource, i'm not a competitor and i'm older than you i think. I'm not bashing. I'm looking for buying a rack but as i was a customer of wholesalecolo, i NEVER got an Epik backup with their Cogent line. That's all so if someone can explain me if it's possible i'll be convinced. All tracert you can make show Epik does not work as a backup if p15-0.core01.mia01.atlas.cogentco.com is down or VERY SLOW which is the same for me...

  11. #11
    Join Date
    Mar 2002
    Location
    Mass
    Posts
    726
    [..]
    9 p6-0.core02.ord01.atlas.cogentco.com (66.28.28.217) 6.561 ms 3.420 ms 2.553 ms
    10 p15-0.core02.ord01.atlas.cogentco.com (66.28.4.62) 4.273 ms 6.893 ms 7.847 ms
    11 p12-0.core01.mci01.atlas.cogentco.com (66.28.4.33) 23.054 ms 23.770 ms 15.856 ms
    12 p5-0.core02.dfw01.atlas.cogentco.com (66.28.4.37) 24.030 ms 39.051 ms 23.874 ms
    13 p15-0.core01.dfw01.atlas.cogentco.com (66.28.4.25) 24.234 ms 26.453 ms 23.883 ms
    14 p13-0.core01.iah01.atlas.cogentco.com (66.28.4.98) 32.931 ms 29.849 ms 49.130 ms
    15 p15-0.core01.tpa01.atlas.cogentco.com (66.28.4.46) 76.417 ms 47.382 ms 49.218 ms
    16 p5-0.core01.mia01.atlas.cogentco.com (66.28.4.57) 53.626 ms 58.391 ms 55.106 ms
    17 g49.ba01.b002802-1.mia01.atlas.cogentco.com (66.28.6.42) 61.524 ms 64.235 ms 56.109 ms
    18 miamfllrh07-custbr-fastduck.ip.epik.net (216.22.64.62) 59.462 ms 60.431 ms 62.776 ms
    19 66.232.5.6 (66.232.5.6) 55.950 ms 59.252 ms 53.690 ms
    20 * * *
    21 * * *
    22 * * *
    23 * * *
    24 * * *
    25 * * *
    26 * * *
    27 * * *
    28 * * *
    29 * * *
    30 * * *
    [....]

    Those times are pretty bad from Verio(what that box is on) to Cogent.
    Jason

  12. #12
    Join Date
    Mar 2002
    Posts
    346
    neverless,

    The server is on Cogent, not Epik.
    The Epik hop you see is a cross connect we have between 2 buildings.

    I can not say much about Wholesalecolo, but us and another company split the clients.
    The bandwidth has also been done correctly.

    If you would like, please e-mail me directly at [email protected], and I can send you an IP on Epik to do a trace on.

    BTW Thanks for everyone trying to PING or trace to serveroutsource.net.
    Our firewall had a busy time. :\

  13. #13
    Join Date
    Jul 2002
    Location
    Vancouver, BC
    Posts
    1,017
    so with epik reselling level3 bandwidth, if i do a trace from within the level3 network to you guys, it should go through epik??

    Tampa, FL to ServerOutSorce - 104ms - http://alice.eng.level3.net/tracerou...routsource.net

    Miami, FL to ServerOutSource - 108ms - http://alice.eng.level3.net/tracerou...routsource.net

    thats an example of cogent routing and their wonderful peering... bouncing around the wee continent just to get to a place in the same city, or atleast damn close by... i mean... epik is a very small network, they should have feeds to level3 closer than cogent does... eh?

  14. #14
    Join Date
    Mar 2002
    Posts
    346
    netdude,

    As mentioned before,

    "The server is on Cogent, not Epik.
    The Epik hop you see is a cross connect we have between 2 buildings.

    I can not say much about Wholesalecolo, but us and another company split the clients.
    The bandwidth has also been done correctly.

    If you would like, please e-mail me directly at [email protected], and I can send you an IP on Epik to do a trace on.

    BTW Thanks for everyone trying to PING or trace to serveroutsource.net.
    Our firewall had a busy time. :\"

  15. #15
    Join Date
    Apr 2001
    Location
    St. Louis, MO
    Posts
    2,508
    BTW Thanks for everyone trying to PING or trace to serveroutsource.net. Our firewall had a busy time.
    If your firewall was busy with just ICMP traffic, you may need a new firewall
    Mike @ Xiolink.com
    http://www.xiolink.com 1-877-4-XIOLINK
    Advanced Managed Microsoft Hosting
    "Your data... always within reach"

  16. #16
    Join Date
    Jul 2002
    Location
    Vancouver, BC
    Posts
    1,017
    if one of these statements are true, the other can't be... ... ... ...

    Originally posted by SrvOutsource
    Epik is the secondary to the CogentCo bandwidth.
    Originally posted by SrvOutsource
    netdude,

    As mentioned before,

    "The server is on Cogent, not Epik.

    ...

  17. #17
    Join Date
    Oct 2000
    Posts
    1,653
    Originally posted by raq4less
    Epik has peering agreements with Level3 and Aparently now Cogent! (That's brand new)
    Wrong. http://www.fixedorbit.com/AS/19/AS19962.htm
    [QuickPacket™] [AS46261]
    Located in Atlanta, GA and Los Angeles, CA
    Dedicated Servers, KVM, Xen & OpenVZ VPS, Co-location, R1Soft Data Backup, Shared & Reseller Hosting

  18. #18
    Join Date
    Mar 2002
    Posts
    346
    netdude,

    Epik is the backup (Secondary) line to Coegnt.
    If Cogent goes down, Epik will take over, which is how our BGP is setup.

    When you do a trace to our server, you should not see Epik, unless Cogent is not working for some reason.

    I do not know how to make this much clearer.

  19. #19
    Join Date
    Jul 2002
    Location
    Vancouver, BC
    Posts
    1,017
    general consensus... that ain't multihoming... heh

    http://www.webhostingtalk.com/showth...threadid=69437

  20. #20
    Join Date
    Mar 2002
    Posts
    346
    Netdude,

    The definition of Multihomed is that the network is on multiple networks, not only one.
    Which the network is one 2 networks, CogentCo and Epik.

    The link you posted is really describing "LOAD BALANCING via BGP4".

    You may want to check on online technical dictionary like at the below, before you start debating something.

    http://searchnetworking.techtarget.c...212611,00.html

    "Multihomed describes a computer host that has multiple IP addresses to connected networks. A multihomed host is physically connected to multiple data links that can be on the same or different networks. "

    or

    http://www.webopedia.com/TERM/M/multihomed.html

    "An adjective used to describe a host that is connected to two or more networks or has two or more network addresses. For example, a network server may be connected to a serial line and a LAN or to multiple LANs."


    Our network IS multihomed according to the official published definition.
    We do not claim it to be load balanced via BGP.

  21. #21
    NetDude,

    You really have absolutely no idea how networks operate nor the concept of IP routing on the Internet, do you?

    "so with epik reselling level3 bandwidth, if i do a trace from within the level3 network to you guys, it should go through epik?? "

    Absolutely not. It would only go through epik if epik were the "preferred" route to the destination. And "preferred" has nothing to do with best path.

    Routes on the Internet are based on a string of AS #'s. The "preferred" route is the string with the fewest AS #'s in it. BGP allows padding from the destination of additional AS #'s so that what may be the shortest path becomes the longest path. That is how traffic meant to flow over a Cogent offering (or any other single carrier offering) is forced to flow over a single carrier offering.

    You could have a server on the Epik network and traceroute to an IP on a multi-homed network with Epik as one of the carriers and it would still route through an alternate carrier if that was the "preferred" path.

    If you are so intent on understanding multi-homing and debating whether certain providers have multi-homed networks, please do your own research 1st and learn a little bit about BGP.

    Shaun Gregory
    CEO
    WholesaleBandwidth, Inc.

  22. #22
    Join Date
    Mar 2002
    Location
    Mass
    Posts
    726
    blah blah blah

    "my network is multi-homed cause it's connected to two networks but data only travels over one if the other one fails"

    blah blah blah

    That's redundent. Technically, if you want to think about it alot, that's correct. But you can run your own ship.

    I learned that in Cisco 5.. advanced routing.

    Whatever floats your boat though, we've been over this before and it's not going to go anywhere.
    Jason

  23. #23
    This really cracks me up, to see someone debating a topic like this. Let me finish this up for you.

    If I have 2 circuits and I have a BGP session with both providers, I can set my local preference on the circuit I want to use HIGHER than the other circuit using a route map.

    The effect of this is that ALL of my outbound bandwidth will go out the circuit that I have set the highest preference to and in the event that my circuit goes down, which in turn causes the BGP session to fail, the other circuit, being the only one left will then have the highest preference and all traffic will flow out that circuit.

    The same thing can be accomplised for inbound traffic by using AS PATH PREPENDS or by asking your upstream provider to adjust their local preference for your netblocks.

    This is not rocket science and a common practice used to provide redundancy while controlling costs.

    Any questions?
    Hosting, VDS, and Colocation Solutions
    by Digital Isles, Inc.
    http://www.digitalisles.com

  24. #24
    BTW, can you do a half cabinet or quarter cabinet configuration?

    Thanks,
    Hosting, VDS, and Colocation Solutions
    by Digital Isles, Inc.
    http://www.digitalisles.com

  25. #25
    Join Date
    Mar 2002
    Posts
    346
    DigitalIsles,

    Sorry, but was that question aimed at ServerOutsource.NET?

    If so, yes we do.
    One of the benifits of having a cage area, instead of individual cabinets.


    David

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •