We now have full support for IPv6 for our dedicated server, Internet transit, and colocation customers. Our IPv6 network utilizes connectivity from nLayer, Hurricane Electric, and Atlantic Metro Communications along with our own local peering agreements. To test out our IPv6 network you can use our looking glass or access our web site, support site, and mirror server which are all IPv6 capable.
All dedicated server and colocation customers would be eligible to receive a /56 (2^72 IPs) in order to assign a /64 (2^64 IPs) per device. Colocation customers would be assigned the entire /56 at no charge, while dedicated server customers would get a /56 allocated on our side and then be assigned a /64 to each device as requested. For dedicated server customers, there would be a one-time $20 fee for each /64 to cover the time needed to setup and configure the IPv6 address space. Colocation and Internet transit customers doing their own switching/routing would be eligible to receive a /48 (2^80 IPs). To request your IPv6 address block, just contact our IPs department by emailing email@example.com or by using the IP Allocation department in the ticket system.
Note: We can offer IPv6 only VLANs for dedicated server, colocation, or Internet transit customers. We can offer reduced pricing over these VLANs to help encourage the overall use of IPv6. Just contact our sales department for details.
This article is relating to our shared hosting platform utilizing Litespeed web server. We no longer offer shared web hosting, but we do offer Litespeed hosting options with our cloud hosting and dedicated server services.
The LiteSpeed web server 4.0 release landed just last week. This new release features, among other things, support for Server Side Includes (SSI), large file delivery (greater than 2 GB files) and mod_security 2.0 support.
To address capacity expansion needs, we've just deployed a new shared hosting web server, web 6. This server has been deployed with the new 4.0 version and it appears to be working well. Our SimGames.net hosting users have been testing the 4.0 version for several weeks, going back to the RC1 release, and everything has been stable and reliable, so, barring any issues, we are hoping that by next week we'll be ready to upgrade across the board to 4.0 on all existing servers. Web 4 will remain Apache and we do not plan to make any changes to that server at this time.
Web 6 is receiving all new accounts effective immediately. If you'd like to try out the new SSI features or other 4.0 related features early and are an existing customer, please contact our support team to have your site moved over to the new server. We'd especially like feedback from any customers currently using SSI on web 4 that are adventurous enough to see if their SSI usage translates well to the LiteSpeed implementation.
We'll post an announcement next week to schedule the cut-over of existing servers.
Lately we've been working hard on getting IPv6 setup network-wide. Getting IPv6 is really more complicated to get setup on the policy and procedure side than it is technically, as technically it isn't all that difficult. The plan will be to allocate a /56 per dedicated hosting/small colocation customers with the plan to be to give each individual server a /64 (the same number of IPs as the entire IPv4 block squared). Larger colocation customers, will be allocated their own /48 for them to distribute as they wish.
Now, this may seem wasteful, but the idea with IPv6 is to reduce the size of routing tables, to make routing easier, etc. by making the first allocation to someone all they will ever need. There should never be a need for an additional/secondary allocation, keeping overall router configurations simple.
For connectivity, we already have IPv6 transit from Hurricane Electric, NLayer, and Atlantic Metro. We also have IPv6 peering with several other networks through the Equinix Exchange and ChIX. We do not have an official release date yet for the IPv6 product, but we hope it does not take much longer. The plan is to initially release it as a "beta" product with no SLA, but that beta period should last no longer than 2-3 months, assuming everything goes smoothly.
This week the American Society for Heating, Refrigerating and Air-conditioning Engineers (ASHRAE) adjusted their recommended upper limit for ambient data center temperatures from 77 degrees to 80.6 degrees. Combining this with reports from Google, Intel, Sun, and HP, you can easily see that servers do not mind "higher" temperatures as much as you would think.
I have to address this many times on tours, as we keep our facility at ~72 degrees Fahrenheit for the inlet temperatures on servers. It seems many people expect the facility to be significantly cooler, say in the 64-66 degree range. As these reports show, keeping a data center at those levels is simply not necessary as it does not prove to have any benefit for reliability, performance, etc. The only real difference is that keeping a facility at those temperatures simply costs a lot more and wastes more energy. It is true, that at warmer temperatures you need to be more wary of hot spots, but with regular monitoring and a properly engineered facility, that should not become an issue.
As a note, our facility even easily met the old ASHRAE standards. You can see in this graph/report that our spec of 72 degrees Fahrenheit and 45% relative humidity are solidly in the area recommended for a Class 1 Operating Environment.
We have just changed our colocation pricing scheme to make it more fair to all of our customers. The bandwidth pricing has not changed, but we have altered the space/power pricing scheme to be more focused on charging for power utilization. This pricing change will not affect existing systems in our facility, but will affect new systems sent to us on our shared colocation plans. We are hoping this change will help move our customers to use more power efficient, and thus more environmentally friendly, equipment by making them take the power usage of their systems into account.
The new pricing is listed as being $35 a month for a 1u server and 0.5A of power, which is a good bit less than what our pricing had been at the base level, $50. This means on any lower power equipment such as switches, firewalls, etc. you'll be looking to pay less. It is then $10 a month per additional 1u of space and then $25 a month per additional 0.5A. That means for a 2u system using 1A of power your base price, before bandwidth, would be $70 a month. This pricing is done per system, not over an array of systems, etc. It should help get reduced pricing for customers using power efficient equipment, while increasing the pricing for those using less efficient equipment. In short, power is our primary cost and restricting factor on these systems, the space itself is cheap.
If you have any questions about the new pricing, please email us at firstname.lastname@example.org