Results 1 to 21 of 21
  1. #1

    Question Looking for 10Mbit/s Colo

    I am looking for a 1U colo with 10Mbit/s unmetered bandwidth.

    ColorQuest.com is the cheapest one I can find so far.
    $40/mo for per U space and $350/mo for the bandwidth.

    Do anyone know some other provider offering a better deal?

  2. #2
    Join Date
    Nov 2003
    Location
    Newport Beach, CA
    Posts
    337
    What region are you looking at?
    Best Wishes,

    Blake L. Smith - blake@xtremebandwidth.com
    XtremeBandwidth.com, Inc. - Wholesale Tier1 Bandwidth!

  3. #3
    Any region in the U.S. is fine.

    Thanks.

  4. #4
    Join Date
    Dec 2002
    Location
    Vancouver, Canada
    Posts
    541
    Egihosting.com's probably the cheapest for your needs.

  5. #5
    Join Date
    Feb 2004
    Posts
    390
    Whatever provider you choose, consider the following:

    1. Bandwidth (Networks)
    2. Support - what support level do they provide? Remote hands? How much is it? etc.

  6. #6
    Join Date
    Feb 2004
    Posts
    390
    from what i heard egihosting has issues with spews and IP's being blacklisted, he.net as a whole has these problems search the forums

  7. #7
    Join Date
    Feb 2004
    Posts
    227
    $350 for 10mbps colo is a good deal in itself.....

  8. #8
    That depends on the type of traffic. I am sure one could easily get 10Mbit/sec of someone who has Cogent gige at that price but that would get them well... cogent.
    www.zubrcom.net | Tel: 1-877-982-7266 / 1-267-298-3232 | sales@zubrcom.net|@zubrcom
    Hosting, VPS, Servers, Unmetered 10, 100 and Gigabit servers - Colocation - Engineering services
    "Elegant solutions to complex problems in the Internet-centric world"

  9. #9
    Join Date
    Feb 2004
    Location
    Louisville, Kentucky
    Posts
    1,083
    Originally posted by zubrcom
    That depends on the type of traffic. I am sure one could easily get 10Mbit/sec of someone who has Cogent gige at that price but that would get them well... cogent.
    Yeah, so, I'm going to post my cogent rant again.

    1) clueless start-up web host buys cogent FastEthernet transit product for $3000/Mo
    2) clueless start-up has no experience with routing, switching, billing, customer service, security, abuse desk, etc. etc.
    3) clueless start-up blames problems on cogent, because that is the party line of third-rate web hosters
    4) cogent acquires bad reputation

    I know you understand all this, because you, yourself, just posted an informed remark about not oversubscribing core or transit in WHT threadid 249143.

    So I'll just ask, have you been a direct customer of Cogent? What specific problems did you have with your service, and why were they unable to resolve them?
    Jeff at Innovative Network Concepts / 212-981-0607 x8579 / AIM: jeffsw6
    Expert IP network consultation and operation at affordable rates
    95th Percentile Explained Rate-Limiting on Cisco IOS switches

  10. #10
    Originally posted by jsw6
    Yeah, so, I'm going to post my cogent rant again.

    1) clueless start-up web host buys cogent FastEthernet transit product for $3000/Mo
    2) clueless start-up has no experience with routing, switching, billing, customer service, security, abuse desk, etc. etc.
    3) clueless start-up blames problems on cogent, because that is the party line of third-rate web hosters
    4) cogent acquires bad reputation

    I know you understand all this, because you, yourself, just posted an informed remark about not oversubscribing core or transit in WHT threadid 249143.

    So I'll just ask, have you been a direct customer of Cogent? What specific problems did you have with your service, and why were they unable to resolve them?
    Yes, I have multiple first hard experiences with Cogent, which ranged from bearable to horrible. I will give them another six month before I even consider looking at them again.
    www.zubrcom.net | Tel: 1-877-982-7266 / 1-267-298-3232 | sales@zubrcom.net|@zubrcom
    Hosting, VPS, Servers, Unmetered 10, 100 and Gigabit servers - Colocation - Engineering services
    "Elegant solutions to complex problems in the Internet-centric world"

  11. #11
    Join Date
    Feb 2004
    Location
    Louisville, Kentucky
    Posts
    1,083
    Again, let's see specifics. When did these "horrible" to "bearable" events happen, and why were they not resolved to your satisfaction?
    Jeff at Innovative Network Concepts / 212-981-0607 x8579 / AIM: jeffsw6
    Expert IP network consultation and operation at affordable rates
    95th Percentile Explained Rate-Limiting on Cisco IOS switches

  12. #12
    Join Date
    Jan 2003
    Posts
    252
    jsw6: Judging by previous posts/arguments he was probably trying to host IRC on it and it got slammed 24/7 with DoS
    Warpline Hosting: Budget SSD Web Hosting, Reseller hosting, VPS Hosting.
    Bitcoin, PerfectMoney, and PayPal accepted.

  13. #13
    Originally posted by jsw6
    Again, let's see specifics. When did these "horrible" to "bearable" events happen, and why were they not resolved to your satisfaction?
    General CS:

    1. Cogent's NOC has no concept of escalation unless they are threatened with cancelation.

    2. Cogent's NOC/CS is incapable of notifying more than one contact in the event of an outage, even if they clearly did not get in touch with the first one and did not leave a message.

    3. Cogent's NOC closes tickets whent he problem had not been resoled.

    Specific

    1. On four separate occasions over a 7 month of service, Cogent's port went from the full duplex on FE to half duplex, their NOC did not notice it. After being told specifically to test for that the person claimed to have done it, when she did not. Three hours later someone did determine that the port on their side went half duplex. This happened several times.

    2. Cogent would sell you full gige or a fully burstable gige when the long haul that feeds it cannot support the traffic, even when from the beginning Cogent's sales rep is told the amount of traffic that could be tossed onto the network. Cogents NOC would refuse to comment on presence of a cogestion on the long haul even after the tickets are opened and escalated.

    3. Cogent is a high-jitter network.

    4. Don't even start me on peer A/peer B and eBGP multihop.
    www.zubrcom.net | Tel: 1-877-982-7266 / 1-267-298-3232 | sales@zubrcom.net|@zubrcom
    Hosting, VPS, Servers, Unmetered 10, 100 and Gigabit servers - Colocation - Engineering services
    "Elegant solutions to complex problems in the Internet-centric world"

  14. #14
    Originally posted by interquad
    jsw6: Judging by previous posts/arguments he was probably trying to host IRC on it and it got slammed 24/7 with DoS
    If 300Mbit/sec towards a single customer who is able and willing to pay for the traffic is killing your network, then you have a bad network design.
    www.zubrcom.net | Tel: 1-877-982-7266 / 1-267-298-3232 | sales@zubrcom.net|@zubrcom
    Hosting, VPS, Servers, Unmetered 10, 100 and Gigabit servers - Colocation - Engineering services
    "Elegant solutions to complex problems in the Internet-centric world"

  15. #15
    Join Date
    Jan 2003
    Posts
    252
    I don't remember saying anything about a 300 mbps attack killing my network. Go back to IRC.
    Warpline Hosting: Budget SSD Web Hosting, Reseller hosting, VPS Hosting.
    Bitcoin, PerfectMoney, and PayPal accepted.

  16. #16
    Join Date
    Feb 2004
    Location
    Louisville, Kentucky
    Posts
    1,083
    Originally posted by zubrcom
    1. Cogent's NOC has no concept of escalation unless they are threatened with cancelation.
    I guess you didn't read the manual. Check out pages 27 and 28 of the Cogent User Guide, which clearly document their escalation procedures.
    http://cogentco.com/Guide/User_Guide_V2.2002.pdf

    2. Cogent's NOC/CS is incapable of notifying more than one contact in the event of an outage, even if they clearly did not get in touch with the first one and did not leave a message.
    You shouldn't rely on your carriers to be an escalation entity for you. This is a job for an inside 24x7 NOC, an answering service, a shared pager, or a phone system capable of ringing multiple parties.

    Further, you shouldn't rely on carriers to call and let you know your transit is down in the first place. That's what clued staff, network monitoring software, and outsource services are for. I don't expect any of my carriers to call me up and say, "By the way, our router failed and your transit will be down pending repair." If they do, that's fine, but the bottom line is it 1) won't cause me signifigant distress to lose a single carrier; and 2) my people are going to be notified, and if I need to do anything about it a member of our staff will call me, not some vendor.

    3. Cogent's NOC closes tickets whent he problem had not been resoled.
    Everyone does that. Be specific. Was it your fault for not following-up on the ticket? Did you ask for filtering or something else Cogent does not offer, and then consider that within your definition of "problem had not been resoled(sic)?"

    1. On four separate occasions over a 7 month of service, Cogent's port went from the full duplex on FE to half duplex, their NOC did not notice it.
    Again, see the Cogent User Guide. It is their policy to configure FastE hand-off equipment for 100/Full in all installations. You say their switch up and changed its duplex setting on its own, four times? I think Cisco would be interested in this tall tale. Perhaps there is a compatibility problem with your ethernet chipset and the chipset used by whichever Cisco device was at your Cogent service location. If auto-negotiation was off on both ends, as is recommended by the user guide, that wouldn't, shouldn't, couldn't happen.

    2. Cogent would sell you full gige or a fully burstable gige when the long haul that feeds it cannot support the traffic, even when from the beginning Cogent's sales rep is told the amount of traffic that could be tossed onto the network. Cogents NOC would refuse to comment on presence of a cogestion on the long haul even after the tickets are opened and escalated.
    The only circumstance under which Cogent has the potential for aggregation congestion is when part of a metro ring fails.

    As to your remark that you were personally able to congest Cogent's oc-192c core with your burstable gige that has been available for less than 3 months, or your full gige which you evidently bought because you couldn't get the right duplex setting on FastEthernet, I am going to call ******** once again. I doubt that the several dozen Cogent customers using their Gige near 100% have had this same experience. Why is it only you? Would you like to re-specify your complaint, and indicate specific peering or other routes?

    3. Cogent is a high-jitter network.
    The Cogent oc-192c backbone, which is designed and implemented congestion-free, is high-jitter? Again perhaps you'd care to cite specific peering, as if Cogent had a jitter problem on their backbone it would be because they have congestion and buffering, or ECMP routes that are of differing latency (which they do not.) I somehow suspect that other direct Cogent customers would have noticed if they had congestion and large packet buffers on their backbone, because that would also introduce packet loss once traffic levels exceeded the amount of transmit time available in real-time outside of a micro-burst setting.

    4. Don't even start me on peer A/peer B and eBGP multihop.
    What is your problem with the way Cogent delivers BGP routes, exactly? It's a design decision and it doesn't cause any signifigant disadvantage!

    Many carriers deliver BGP routes via a multihop session. AboveNet does so in almost all circumstance. Will you now begin to bash AboveNet engineering practices on the basis that they utilize ebgp multihop for customer BGP sessions?

    Cogent does this because they set out to offer a uniquely priced and delivered product, and they had some cost vs engineering decisions to make. Did you know that Cogent will not sell more than 20 FastEthernet ports on one edge router, because it only has 2Gb/sec of uplink capacity? I would rather them make that trade-off, than deploy a router capable of delivering full BGP routes in every building just to avoid ebgp multihop. Such minutia!
    Last edited by jsw6; 03-18-2004 at 03:38 AM.
    Jeff at Innovative Network Concepts / 212-981-0607 x8579 / AIM: jeffsw6
    Expert IP network consultation and operation at affordable rates
    95th Percentile Explained Rate-Limiting on Cisco IOS switches

  17. #17
    Join Date
    Dec 2002
    Location
    acton, MASS
    Posts
    229
    4. Don't even start me on peer A/peer B and eBGP multihop.
    what is wrong with this setup? it's called buying high-port density L3 switches while delivering unblocked fast performance for the customer at low cost.

    just fyi, our bgp session with even the "traditional tier-1 (Wooo)" carriers have ebgp multihop 2 for varying reasons. any clued noc admin should have no problem with cogent's setup.

    -J
    TowardEX Technologies, Inc.
    Metro-Boston Affordable Colocation, Dedicated Servers & IP Consulting
    http://www.towardex.com
    inquiry@towardex.com

  18. #18
    Join Date
    Feb 2004
    Posts
    390
    this thread has been officially hijacked

  19. #19
    Originally posted by zubrcom
    If 300Mbit/sec towards a single customer who is able and willing to pay for the traffic is killing your network, then you have a bad network design.
    how much pps?

    paul
    * Rusko Enterprises LLC - Upgrade to 100% uptime today!
    * Premium NYC collocation and custom dedicated servers
    call 1-877-MY-RUSKO or paul [at] rusko.us

    dedicated servers, collocation, load balanced and high availability clusters

  20. #20
    Join Date
    Dec 2002
    Location
    acton, MASS
    Posts
    229
    Originally posted by zubrcom
    If 300Mbit/sec towards a single customer who is able and willing to pay for the traffic is killing your network, then you have a bad network design.
    i see. let me enlighten you on at what point networks break during dos attacks from 5-years service provider consultation experience.

    it doesn't happen at bits/sec when you have uber pipes like gig-e. it happens at Packets Per Second, or pps.

    so how many pps? you sure it was not your 20 dollars netgear switch bought off ebay dropping packets that cogent was sending you at full line rate?

    -J
    TowardEX Technologies, Inc.
    Metro-Boston Affordable Colocation, Dedicated Servers & IP Consulting
    http://www.towardex.com
    inquiry@towardex.com

  21. #21
    Originally posted by jsw6
    I guess you didn't read the manual. Check out pages 27 and 28 of the Cogent User Guide, which clearly document their escalation procedures.
    http://cogentco.com/Guide/User_Guide_V2.2002.pdf

    When a service is experiencing 2-10% packet loss two hops past the customer port in provider's network and the ticket had been opened with providers noc for over 24 hours, if the customer calls the NOC and demands escalation, NOC should escalate. Of course, instead one may look for a procedure to toss at a customer.

    It is called customer service, or lack of there of.

    You shouldn't rely on your carriers to be an escalation entity for you. This is a job for an inside 24x7 NOC, an answering service, a shared pager, or a phone system capable of ringing multiple parties.
    If a carrier cannot do it, then it should not promise that as it attempts to sell the service. If it is not a service that carrier promises to provide, then it should not allow its sales people to promise it.

    When carriers ticketing system of cannot handle "call A, call B and call C in an event of a ticket opened", it reflects on customers perception of carriers service.

    When the carrier keeps insisting that it had been a snafu and they can in fact handle more than one notification contact on the account, and then fail to do that again, then it again reflects very poorly on carriers ability to service customers.


    Further, you shouldn't rely on carriers to call and let you know your transit is down in the first place. That's what clued staff, network monitoring software, and outsource services are for. I don't expect any of my carriers to call me up and say, "By the way, our router failed and your transit will be down pending repair." If they do, that's fine, but the bottom line is it 1) won't cause me signifigant distress to lose a single carrier; and 2) my people are going to be notified, and if I need to do anything about it a member of our staff will call me, not some vendor.
    If as a part of a sale of a car you are selling me roadside assistance, I would be stupid not to call on it when I am stuck.
    The argument "why the hell do you rely on our roadside assistance? On page 40 of our users manual we say should call AAA" would sound rather funny.

    Let me correct it, the transit was not down - Cogent was experiencing nasty packet loss, after five minutes of demonstrating that packet loss, it was shifted on the provider that did not have the issue.


    Everyone does that. Be specific. Was it your fault for not following-up on the ticket? Did you ask for filtering or something else Cogent does not offer, and then consider that within your definition of "problem had not been resoled(sic)?"
    No, I asked for RFO. The same way as I asked for one from Verio, AT&T, Level3, and others. Oh right, I forgot to check the User Guide. Do I need to send a properly formatted RFO request to a specific person?


    Again, see the Cogent User Guide. It is their policy to configure FastE hand-off equipment for 100/Full in all installations. You say their switch up and changed its duplex setting on its own, four times? I think Cisco would be interested in this tall tale. Perhaps there is a compatibility problem with your ethernet chipset and the chipset used by whichever Cisco device was at your Cogent service location. If auto-negotiation was off on both ends, as is recommended by the user guide, that wouldn't, shouldn't, couldn't happen.
    The other side was locked in 100/Full. The fun part was that the other side was showing the port configured 100/Full and Actual 100/Half. The same switch carried a couple of other providers with the identical setup. One of those providers had exactly the same gear that Cogent had on their end and it never failed on them. Oh, and we relocated Cogent onto a different port and put the provider that did not have a problen on the one that was connected to Cogent. The problem migrated together with the Cogent's connection.

    Of course the other provider could notify more than 1 number in an event of the problem, the NOC would call us if they saw somethinig whacky, and the NOC manager was extremely helpful in making sure that we were happy with any explanation of any event that affected our service. Of course it was that strange company called Verio, which has this proposterous idea that support is what keeps the customers in the event of a problems.


    As to your remark that you were personally able to congest Cogent's oc-192c core with your burstable gige that has been available for less than 3 months, or your full gige which you evidently bought because you couldn't get the right duplex setting on FastEthernet, I am going to call ******** once again. I doubt that the several dozen Cogent customers using their Gige near 100% have had this same experience. Why is it only you? Would you like to re-specify your complaint, and indicate specific peering or other routes?
    Dunno, I guess maybe cause when that service was accepted there, the entire super cool super fast super oc102c backbone was not active there? Failed, taken down for maintenance, etc, I do not know, and frankly, do not care to know, since NOC insisted that there had not been a problem to begin with anyway. I


    The Cogent oc-192c backbone, which is designed and implemented congestion-free, is high-jitter? Again perhaps you'd care to cite specific peering, as if Cogent had a jitter problem on their backbone it would be because they have congestion and buffering, or ECMP routes that are of differing latency (which they do not.) I somehow suspect that other direct Cogent customers would have noticed if they had congestion and large packet buffers on their backbone, because that would also introduce packet loss once traffic levels exceeded the amount of transmit time available in real-time outside of a micro-burst setting.
    I tend not to pay full transit price when I want someone to take my packets only to their network. It seems to me that their interconnects to a lot of places are rather congested. However, let me encourage everyone here to pay full transit price to access the carrier's customers with low latency, low jitter and no packet loss.


    What is your problem with the way Cogent delivers BGP routes, exactly? It's a design decision and it doesn't cause any signifigant disadvantage!
    I like the idea of keeping it simple, and stupid. I do not like to maintain multiple prefix lists for the purpose of doing the same thing, or multiple route-maps for 'announce-me-normally'. I like the idea of knowing that I just need to worry about my ability to handoff traffic directly to whatever router I am directly connected to. I am such a lazy bastard.

    [QUTOE]
    Many carriers deliver BGP routes via a multihop session. AboveNet does so in almost all circumstance. Will you now begin to bash AboveNet engineering practices on the basis that they utilize ebgp multihop for customer BGP sessions?
    [/QUOTE]

    The funny part is that when one asks most of those carriers for ebgp multihop to feed provider's view of the net into the software that later showed up as SockEye or RouteScience, the same carriers would not do it.


    Cogent does this because they set out to offer a uniquely priced and delivered product, and they had some cost vs engineering decisions to make. Did you know that Cogent will not sell more than 20 FastEthernet ports on one edge router, because it only has 2Gb/sec of uplink capacity? I would rather them make that trade-off, than deploy a router capable of delivering full BGP routes in every building just to avoid ebgp multihop. Such minutia!
    Right, and that's why when you buy Cogent, you get Cogent and all the beauties uniquely associated with it. I encourage my competitors to use it.
    www.zubrcom.net | Tel: 1-877-982-7266 / 1-267-298-3232 | sales@zubrcom.net|@zubrcom
    Hosting, VPS, Servers, Unmetered 10, 100 and Gigabit servers - Colocation - Engineering services
    "Elegant solutions to complex problems in the Internet-centric world"

Posting Permissions

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