Server Advise 130,000 Page Views Day

Hosting
1

Currently have a wordpress site project which is hosted and managed in the US.

The sites visitors are all New Zealand based.

Is a linux cpanel vps server, has 1.5gb ram
using 38% ram on average, average load 1.35 (16cpus shared)

Traffic:
17,000 visits per day
24,000 page views a day

This site has a high page load 1.2mb each page(lots of images, maps, functions etc)
161 cpu requests per page (including external).

Plaform:wordpress
Cache:Wp-supercache
Webserver:Lightspeed
CDN:MAX CDN
plugins:21 (all required)

Within two months paid traffic will increase to 100,000 visits a day (roughly 130,000 page views a day)

Will i need to go to a dedicated? and what specs would you recommend

This needs to be fully managed and secured by the host 99.999% uptime business critical

There is a Hybrid vps plan available with guaranteed 4cores and 10gb ram will this surfice?

Is a 100mbps port ok to run this?

I’m asking as my current host is unable to give me an estimation without trying the plan.

Thanks for your help

2

pc,

Please check your requirements, especially the amount of RAM.

I run both NZ and US websites out of servers in the US. Fast loading times since Hamilton stopped being such a bottleneck!

Regards,

DK

3

For that level of Page Views you want 1, maybe 2 dedicated servers (1 web, 1 database) with fast disks - SAS RAID10 recommended. You will start to run into database performance issues with that level of PV, most won’t support the heavy I/O that it will have.

Matt

4

cough Cloud hosting (scales with need), Amazon, Microsoft Windows Azure both offer fantastic cloud hosting as well as full CDNs.

5

You mean the same Amazon that consistently has major failures every quarter?

6

I wouldn’t know, I don’t personally use Amazon’s services in that regard. I’ve been using Azure, personally.

7

Hi Matt,

As you well know (running a hosting company), no service can guarantee 100% uptime.

As people try to save money using cloud service by using inexpensive, low latency network connectivity between zones; however although this saves money it is an inherent risk as opposed to people duplicating their web sites in each region. This zone based hosting comes with some types of access problems. DB’s for example, can be really bad if their stuck in different zones; not to mention the ‘what happens if a hurricane hits region X,Y, X’. When networking, hardware routing, and or too much mirroring bandwidth clog up zones this approach can be less than 100%. It still however offers people a way to do it so they can afford it, that they otherwise may not if choosing a traditional host (in all of their worldly locations).

Non network connectivity between zones types of hosting should always be more stable, but doesn’t offer the same promise of spreading the latency (at a low cost) between the zones.

It kills me as marketers branded ‘Cloud Hosting’ and many seemed to forget the fact that there is still networking, blade servers, cabiling and possibly clustering technologies and wireless technologies that have the same potential to fail as they always did. All cloud has allowed us to do is to spread around risk, but it doesn’t eliminate it.

I’m certain that your service runs well and that you do very much know what your are talking about. My only point is that there may be different reasons that someone scales up a service in Amazon or Windows Azure as they may be happy with the pricing and the service most of the time and can deal with the types of service outages come with the territory. Otherwise someone should look at hosting in a less complicated environment that has a better chance of overall stability.

Regards,
Steve