Virtualware
Example cost: Virtual Private Cloud (updated)
Feb 10th
Back in September, 2009, I had written a post with a quick overview of what a private cloud (or infrastructure) looks like and some basic costs and information, including why it is a great product (I am biased).
Since then, Dell has retired the PE2900III model server and items change, this is an update for the basic configuration.
Reminder graphic:
2 Servers, 1 firewall
So, originally, the physical servers were configured as:
- Dell PE2900III (reasonably priced, very reliable, I have spares on the shelf)
- 4 ethernet ports (2 built in, 2 port card installed, more can be added)
- 2 73GB SAS drives mirrored together for booting VMware vSphere 4
- 32GB RAM (48GB is max for this hardware platform)
New servers look like:
- Dell PET610 (reasonable price, very reliable, spares to go onto the shelf)
- 4 ethernet ports (all built in)
- 2 80GB SATA drives mirrored together for booting VMware vSphere 4
- 48GB RAM (192GB max available – very expensive)
The reason for the RAM change is that I am seeing a 2:1 (or higher) ratio of RAM to CPU usage in terms of percentage, and 48GB is a good place for this sized system. Also, the newer Xeon 55xx series processors uses RAM sticks in 3s instead of 2 or 4 at a time. 48GB is 12 4GB sticks of RAM. The newer 55xx series of processors also has working hyper-threading (or H/T) and I am seeing very nice performance on servers deployed using this processor family in our network.
Cost difference? The original posting listed had estimated the cost at $1,600.00 per month (see previous post), and I estimate this to be very close, inching up to approximately $1,700.00 per month, and this number should be high. (for accurate pricing, please contact ipHouse sales people, they can run up a quote based on real numbers)
Over Subscription vs Over Capacity – huh?
Jan 15th
Recently, a whole slew of tweets showed up across my feed dealing with the perceived and measured issues across multiple public cloud providers infrastructure.
Image via Wikipedia
One of the posts comes from Chris Hoff (this post in particular) that describes quite clearly what the differences are.
Service providers (anyone doing cloud services, virtualization, colocation, bandwidth, whatever) live upon the idea of over subscription. We make our revenue banking on the fact that not everyone needs their full allotment all the time.
With over subscription there is a chance of reaching a state of over capacity. Anyone using this business model needs to understand that they must be ready for it to happen. It isn’t an issue of ‘will’ but an issue of ‘when’. Good engineering can keep the ‘when’ at bay, virtually forever, and that is what you need to be prepared for.
Another recent posting via The Register (featuring @GeorgeReese) had some data dealing with network latency within the Amazon EC2 network. I don’t have any opinions about what was in this article, but it is something that is going to come up again and again as this new model of computing (for the masses) solidifies and grows up on the Internet.
Comments?
![Reblog this post [with Zemanta]](http://img.zemanta.com/reblog_c.png?x-id=e4ff4cfa-35d8-4078-9c3d-e929e6b33a9d)
VMware releases vSphere 4.0 Update 1
Nov 20th
The crowd cheers!
Though that might just be me :)
I am very excited for this update and I hope that it goes as smoothly as other updates have gone.
I have already updated 2 non-production systems (both ESX) and I am in the process of updating some non-production ESXi servers.
One of the items I need to schedule is the update of the vCenter software – this will require an actual outage for customers who are managing their VMs through ipHouse.
I hope that this update fixes 2 problems I have seen in the past (from the vSphere Client point of view):
- Inability for a datacenter administrator to view or clear host alerts and alarms from the hardware
- Interesting permissions issues that end up being more restrictive than the topic states
Both of these have been difficult to work around.
The first one requires the customer to contact us to ‘reset’ their hardware sensors.
The second one has actually hampered one customer from controlling their VMware cluster because of issues dealing with datastore management and the ability to attach an ISO image to their newly created VMs.
One item that is really exciting is the raising of the number of vCPUs (virtual CPUs) per physical CPU core to 25 per. Since we are able to sell VMware based virtual servers to customers, the ability to scale this higher could mean higher savings long term. Be nice to update pricing to reflect these savings later as well, though don’t know if there will be enough right now. At this point we really aren’t reaching the previous limit of 20 vCPU per pCPU.
Added support for Microsoft Windows Server 2008R2 is also welcome as we were wanting to deploy this version of Windows for customers.
Full release notes can be found here and are worth reading if you are into the enterprise products from VMware.
Related articles by Zemanta
- VMware boasts a half mil vSphere downloads (theregister.co.uk)
![Reblog this post [with Zemanta]](http://img.zemanta.com/reblog_c.png?x-id=fdcb7eed-284b-426e-a61b-8138e593b7f6)
Example cost: Virtual Private Cloud on VMware
Sep 4th
In the last 6 months, I have helped multiple customers achieve their dream of a virtual machine environment built for them exclusively, but with abilities to control their virtual machine setup, configuration, turn up, tear down, etc. These dedicated infrastructure environments are in the ipHouse data center.
This isn’t ‘cloud computing’ as many people think of it (thanks to Amazon EC2 and the like), but it is pretty close to that vague definition, and with far more control available in terms of everything-vm-wise.
What do I mean? With this virtual private cloud, a customer can set up 3 Ubuntu systems, 2 Windows Server 2003, 1 FreeBSD, and 7 Windows Server 2008 systems. There really isn’t anything novel about this (again, reference Amazon EC2 and the like).
What is novel is that the customer can configure these VMs as they wish. Disk space allocation, partitioning, memory configurations, number of vCPUs. Basically, if you can do it on a physical server – you can do it virtually.
Another differentiating feature is that VMware vSphere 4 supports many operating systems while most public cloud providers offer a very limited number in comparison. This choice alone can be enough to warrant looking at this kind of solution.
No per hour fees, no storage fees (above what the customer has purchased), highly available (if configured to do so), dynamic resource scheduling (if configured to do so), bandwidth fees that are predictable. (see VMare vMotion and Storage vMotion, VMware HA, VMware DRS via website)
I’ll build a configuration example offering shared storage between the VMware physical servers. I’ll be doing some cost estimates for the per month fees. These estimates will be high and are purely shown for example. You would want to contact ipHouse Sales to get a real idea for the costs involved.
VMware releases HCL for ESXi 4.0
Jun 8th
Whew! And about perfect timing for me.
I was helping a friend understand that he could not operate ESXi 4.0 on his old server because of the requirement for a 64bit CPU environment.
Today, VMware released this blog post touting the updates and a link to their actual hardware compatibility list, which covers vSphere 4.0 ESX and ESXi, and I bet it still has the older 3.5 HCL as well.
That’s a lot of servers… thanks VMware!
