amuck-landowner

Load Balanced Cluster •• From $3.71/mo •• Dedicated IP •• SSAE-16 II •• Since 1995!

datarealm

New Member
Verified Provider
Offering web hosting solutions since 1995!

======================

We have a wide range of services at unbeatable prices hosted from our Phoenix, AZ SSAE-16 Type II Certified Data Center. We've been hosting websites for almost 20 years and put all of that experience and knowledge into our Network and our Datacenter to ensure the best service possible.

This is NOT your typical shared hosting setup!  Our load balancing gives you:

- maximum site performance (great for wordpress, eCommerce, etc!)

- maximum site uptime

Most hosting companies place your site on a server with 200 of your closest friends.  When that server gets overloaded they wait for you to complain and then migrate the site to a less loaded machine until you complain again as users are added that system.  Likewise if the single server crashes, your site is offline until they correct the issue.

Datarealm's load balancing solves both of these issues!

Our Starter Account, from as low as $3.71/month includes:

Unlimited Databases

Unlimited Email Addresses

WordPress Easy Installer

Multiple scripting language support (eg, PHP, Perl, etc)

Clustered and Load Balanced

50 GB Premium Bandwidth

3000 MB Disk Space

======================

ORDER NOW

Want to test our network?  Try our new LookingGlass:

http://lg.datarealm.com

Our datacenters have all the usual trimmings (UPS failover to generator, etc) as well as fiber links from a huge array of providers.

For more details about us, please visit:

http://www.datarealm.com

 
 

drmike

100% Tier-1 Gogent
Most hosting companies place your site on a server with 200 of your closest friends

Shared hosting, HAHAHA!  No offense, but 200 is a container count for OpenVZ on same server.

Share hosting, pfft, thousands of random people on a server, especially today with big CPU, huge RAM, SSD, etc.

I'd like more info about this load balancing and failover aspect and how it gracefully deals with such for an end customer.  Sounds nifty.
 

datarealm

New Member
Verified Provider
We have clusters of machines (one cluster for http/https, one cluster for pop, one cluster for smtp) sitting in front of shared storage.   Request comes into the load balancer and using a weighted least connection algorithm is sent to a physical box for handling.  The reply comes straight out from the physical box so there is no bottleneck at the load balancer.

All services are health checked every few seconds and dropped from the load balancer when they fail, and are placed back in when they are restored.  Of course we can manually drop systems so that we can do updates, upgrades, etc on physical nodes without service interruption.  Likewise we can scale up by simply introducing a new physical node to the load balancer.

We first implemented this setup back in 97 or 98.   It's worked pretty well thus far. :)
 
Last edited by a moderator:
Top
amuck-landowner