Managed Hosting Using Infrastructure-as-a-Service

Frequently Asked Questions

Absolutely not. Once a physical server is virtualized it is rendered into a series of flat files and these can easily be moved either to either a local server or to another provider if you ever decide that this isn’t the right fit. Virtualized servers can even be converted back to physical servers.

What’s the big deal about virtualization and why do I care?

In the physical world, we have typically maintained a one to one relationship between an application and the underlying hardware, an example of which would be an email server or database server. We do this to minimize the possibilities of instability and incompatibilities caused by multiple applications.

While this worked well enough in the early years, two factors have changed that have made this model impractical. The first is that hardware has become very powerful to the point that most servers spent their entire lives at less the 5% utilization which is a waste of money, electrical power and cooling costs. The second is that more and more applications are being deployed within companies (CRM, Content Management, Document Management, Unified Communications, are but a few) and consequently we end up with “server sprawl”. It is not atypical to find even small companies with a minimum of 3 to 5 servers today.

Clearly something needs to change - and it has. Virtualization allows us to partition a physical server to run multiple guest operating systems simultaneously. We have seen server consolidation ratios as high as 10:1 and now see utilization rates of physical servers back up to 80%. The cost savings from the hardware alone is substantial; once we factor in decreased management costs, decreased power consumption and cooling costs the argument for virtualization becomes very compelling.

You talk about virtualization providing agility and flexibility. What does this mean?

The flexibility and agility comes from the fact that once a server has been virtualized, we can very easily clone it to make, for example, a copy or run a snapshot prior to running a risky upgrade and revert back to it in the event that it goes poorly. These are significant benefits to any organization. We can also deploy new servers from standardized templates in a matter of hours rather than the weeks it would take to order a box, rack it then build and deploy it.

You talk about high availability for my servers, what do you mean?

The virtualization product we use is the most advanced product on the market and is called VMware vSphere 4. It provides the ability to monitor any individual VM and in the event that the underlying host that is running it fails, it can spin up a new copy of it on another host, generally with a minute of the failure. It can do this because the VM’s files don’t reside on the individual host servers but rather on shared storage that is common to all the hosts. This fully protects your virtual servers from any hardware failure.

OK, housing my severs in the data centre sounds great but what if I lose my connection to the data centre? Won’t I lose access to all my systems?

In the unlikely event that your connection to the data center goes down, yes you would lose access to the systems. However the 10Mbps and 100Mbps connections that we provision come with a maximum 4 hour service level agreement (SLA) so, in the unlikely event that you do experience an outage, the expectation is that it would be of a short duration. Historically, these connections have been very stable and we expect them to only get better.

Additionally, some clients like the option of keeping one small file and print server local to the office. This is also a popular choice for those clients with the smaller 10Mbps connections to cut down on the overall network traffic even when the connection is stable. This small local server keeps a replicated copy of the flat files (word, excel, etc) so that access for these files remains local, but backups are made to the data centre. Additionally print jobs are kept local as well.

How are backups managed?

There are a couple of ways that backups are done based on the requirements of the client. The first thing to note is that once your servers are virtualized, we can easily take snapshots of them on a scheduled basis as a first line of defense. Snapshots allow us to go back to a point in time and are a great way to mitigate against problems arising from risky system patches or application updates. We also can continue to do conventional client based backups from within the guest. Either way, the backups get written out to tape on a daily basis to our automated tape library system and validated on a daily basis.

Is this solution “unique” to MikeTango Group? Did you invent this, or is this available elsewhere?

“Infrastructure‐as‐a‐Service” (“IaaS”) is becoming more commonplace in the US and is typically being targeted at medium‐sized companies. Most of the emerging IaaS providers have elected to provide only the infrastructure, leaving the configuration and management of the systems to the individual clients. This does not work for our clients, as they typically don’t have the staff or expertise to do this. Also, many of these providers don’t support Windows-based systems and most do not support VPN access to their cloud so they can’t be tied back into your existing environment. Consequently these offerings are not appropriate for our clients. Interestingly, some telco’s such as Primus have just started to offer exactly this type of a model, with the same shortcomings noted here.

What MikeTango is offering is a fully featured IaaS complete with the same full range of managed services we have always offered our clients. We don’t intend to offer our IaaS service to any customer who we are not also supporting on an on‐going basis.

We believe the bottom line hasn’t changed – MikeTango continues to meet your overall business needs with appropriate technology solutions, at as an effective cost as possible.