[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Colo in Africa
- Subject: Colo in Africa
- From: akshay at mongodb.com (Akshay Kumar)
- Date: Tue, 16 Jul 2019 18:53:44 +0100
- In-reply-to: <CAMhuimhEqzhR-b-56frE=UJpPabAexOZb8YivF9__CRyd765rg@mail.gmail.com>
- References: <CAMhuimgGz+hRf4zQpwEb3o+3-hih75=zXdDgv5dU6cdVZ4-5fA@mail.gmail.com> <CADrHqgyk3akd0vai3W3VWS3zPn7wp=Qr0vM9MtmcqfWuDh97YA@mail.gmail.com> <CAMhuimh7LZLpzoL8Js7wzHhxzgPr3thXjMOCZoBg6KXEJ2RodA@mail.gmail.com> <CADrHqgxWETbE6UY+AOFK-eKZOcs=Raz71_YbYYjq8ydc7twwrA@mail.gmail.com> <CAMhuimhEqzhR-b-56frE=UJpPabAexOZb8YivF9__CRyd765rg@mail.gmail.com>
Then you are "doing it wrong(tm). Good luck.
On Tue, Jul 16, 2019 at 5:40 PM Ken Gilmour <ken.gilmour at gmail.com> wrote:
> These are actual real problems we face. thousands of customers load and
> reload TBs of data every few seconds on their dashboards. We have busy
> servers. We tried cloud. I passionately hate it. We choose to use Bare
> Metal.
>
> On Tue, 16 Jul 2019 at 10:34, Akshay Kumar <akshay at mongodb.com> wrote:
>
>> Go look at the actual specifications for one of the metal boxes - you are
>> not going to come close to maxing anything out with the workload you
>> describe. FSB hasn't been a thing in over a decade. If you really wanted to
>> go crazy you could do some build a custom solution in FPGA on the F1s.
>>
>> It's a moot point since none of this is going to be available in time but
>> perf is a bogus reason and a lot of the times price is too.
>>
>> On Tue, Jul 16, 2019 at 5:12 PM Ken Gilmour <ken.gilmour at gmail.com>
>> wrote:
>>
>>> Speed is not the issue, it's IO. Also streaming 100Gbps of video is very
>>> different to streaming 100Gbps of files smaller than 100kb (average of
>>> about 30kb) the issue on the network level is the number of connections and
>>> CPU, on the server side it's IO and FSB
>>>
>>> On Tue, 16 Jul 2019 at 08:55, Akshay Kumar <akshay at mongodb.com> wrote:
>>>
>>>> The 2nd requirement seems artificial. The new hypervisors have come a
>>>> long way and the overhead is minimal. Also you can run bare metal instances
>>>> in AWS if you really need them with 100Gbps.
>>>>
>>>> Just just use the South Africa AWS region.
>>>>
>>>> On Tue, Jul 16, 2019 at 3:35 PM Ken Gilmour <ken.gilmour at gmail.com>
>>>> wrote:
>>>>
>>>>> Hi Folks,
>>>>>
>>>>> I work for a Security Analytics org and we're looking to build a small
>>>>> POP in Africa. I am pretty clueless about the region so I was wondering if
>>>>> you could help guide me in the right direction for research?
>>>>>
>>>>> The challenges:
>>>>>
>>>>> 1. Network needs to be able to receive millions of small PPS (as
>>>>> opposed to serving smaller numbers of larger files).
>>>>> 2. Can't be cloud (need bare metal servers / colo). We use the
>>>>> full capacity of each server, all the time.
>>>>> 3. Must have good connectivity to most of the rest of Africa
>>>>> 4. We can initially only have one POP
>>>>>
>>>>> This is not like a normal website that we can just host on "any old
>>>>> provider", the requirements are very different.
>>>>>
>>>>> Is there a good location where we could either rent bare metal servers
>>>>> (something like Internap - preferred) or colocate servers within Africa
>>>>> that can serve most of the region?
>>>>>
>>>>> "Good" is defined as an area with stable connectivity and power, no
>>>>> legal restrictions on things like encryption, and good latency (sub 100ms)
>>>>> to the rest of Africa.
>>>>>
>>>>> Our two closest POPs are in Singapore and The Netherlands, so I'd like
>>>>> something closer to the middle that can serve the rest of Africa. Middle
>>>>> East will be deployed after Africa.
>>>>>
>>>>> I hope this is the right place to ask.
>>>>>
>>>>> Thanks!
>>>>>
>>>>> Ken
>>>>>
>>>>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.nanog.org/pipermail/nanog/attachments/20190716/5cf8346d/attachment.html>