<Theme tune>On Sneaky Peak previously, Real-Time Analytics saved the day, while MSM (Multi-Site Manager) made global configuration management a snip. This time we have the wow factor...</theme tune>
As you trawl the Interwebs reading about all this Cloud stuff, it swiftly becomes apparent that there is more than one school of thought on the benefits (or not) to an organization. What all these contrary opinions have managed to do is split the potential users of “The Cloud” into a number of groups cited below.
- The 100 Percenters – these are (typically) the start-up application development company, these guys can stick everything in the cloud. Start small, develop in an agile way and scale as required.
- The number crunchers – there are a number of organizations (pharmaceutical companies are a good example), where there is a requirement for a huge amount of compute resource to do a specific task, in an on demand type of way.
- The hybridizers – this last group is pretty much everyone else. This is where you already have investment in a datacenter, have made a bit of progress towards virtualization and now would like another arrow in your quiver.
Zeus can typically assist groups 1 and 3 mentioned above, but particularly group 3. The capability of spinning up a VM (Virtual Machine) or two in any number of cloud infrastructures has now been around for a long time (in I.T. terms), and the tools for managing these infrastructures are getting there. However, the piece that has been missing is the functionality that delivers your users to the correct server, in the correct location, based on a set of business criteria.
If for example, the connection is a credit card transaction you would most likely prefer to deliver this from your own physical datacenter. However, if this datacenter is busy you may want to deliver it from a virtual environment you have sitting in a co-location facility. Finally, if your datacenter is really busy and you have a lot of requests coming in for your static content, you may want to have your cloud deployed servers handle these transactions, which can scale to cope with the demand auto-magically.
The above is operational Nirvana, and it has been next to impossible to achieve this without an awful lot of effort on the part of the staff responsible for the delivery of the application.
Dramatic pause...
That is, until now!
This is really what the bulk of advancements in Zeus v7.0 are all about, solving the above issue for organizations. The advancements in Zeus v7.0 will allow organizations to deliver their users to the best server, in the best location, for the best service, under present conditions. There are a number of features that combine to make this all happen. The main (and new ones) are the MSM (Multi-Site Manager) that makes the management of these locations possible, the integration of our GLB (Global Load Balancer) into the MSM, and the new Auto-Scaling feature of ZTM (Zeus Traffic Manager).
So, taking our example above again, and looking at the details we see that it is the MSM and GLB that direct the users to the different locations (in this case based on loading at the locations). As the main location gets busy then more traffic is directed to the other locations (applying rules about the content types). The big difference at this point is then what happens when our cloud or virtual environments get too busy? The new Auto-Scaling module now has a part to play, it has thresholds of acceptable performance configured, and some instructions to how many virtual machines can be running. The Auto-Scaling module takes the decision when the thresholds are crossed about what to do. If the module concludes that another virtual machine is required then it communicates with the virtual or cloud infrastructure’s API to provision a new virtual machine. Once this is up and running it is added to the relevant pool and traffic is sent to it (relieving the other virtual machines that are the members of the pool).
When performance issues slacken off, then the ZTM is able to decommission virtual machines down to the configured minimum number, and waits for the next peak of traffic. Zeus Magic!
Check in next week for Part 4.
Comments