I was recently on an island and it got me thinking about whether a set of close islands can support a highly mobile cloud? If not what would be needed to make the Islands Cloud safer from the vagaries of Mother Nature, such as hurricanes, volcanoes, and earthquakes. Can a cloud provider be based on an island? or would it need to be on every island? Only the mainland?

A cloud provider needs several things:

  • Large footprint buildings with adequate cooling, electricity, etc.
  • Fast interconnects between its various locations
  • Redundancy in the form of building, functionality, etc.
  • Protections from all forms of disasters
  • Treaties

Each of these areas are very important, with some of these being almost impossible to implement on just a single island, but could be implemented across multiple islands, if you are willing to invest in the network and power capacity required. For example, islands tend to be in the path of typhoons, hurricanes, or tsunamis. Given this, protections from all forms of disasters take on new meaning that you would not get in a land locked part of a country.

Network bandwidth also becomes a major issue as cables to and from an island need to be laid on the ocean floor and extend many miles and cost quite a bit to put in. We could look at satellite but many things can affect satellite reception and while fast, it is not fast enough for high speed migrations.

A key component of a cloud is its agility. Eventually, as was demonstrated at VMworld, it will be possible to have agile virtual machines within a cloud that not only can move between hosts within the same datacenter but move between hosts amoungst multiple datacenters on different nearby islands. Given enough bandwidth this would be a boon to any cloud over an island. If you are having problems on island A, your data could be moved seamlessly to island B, or even the mainland. This would require the use of long distance vMotion/LiveMigration technologies.

If this is an island nation and your cloud infrastructure centers are global then the cloud provider and a customer will need to worry about whether or not it is legal to have their data within another country if there is a disaster pending. If not, then what is the backup or disaster recovery plan. This may require paying close attention to treaties and other global considerations, as having this level of agility has its own legal and emotional concerns.

Another item to consider is producing disaster recovery backups, who is responsible, what if the cloud provider goes under, cannot restore quickly due to the foulness of the disaster, where is your data then? This is no different than any other disaster recovery, the data owner is responsible for backing up and safe-guarding this data. You can use the same long-distance LiveMigration capabilities to get a live copy of your own data. If necessary, you float it on the ocean, setup a laptop with a satellite link and you are back in business.

These capabilities could lead to smaller footprint buildings using less electricity and improved protections from all forms of disasters as well as improved automation and agility of the customers data.

Now if Microsoft, Terremark, Amazon, Hosting.com, and the other cloud providers can get together so that VMs could be moved easily between their disparate products (Hyper-V vs Xen vs vSphere) quickly and without much downtime, then we would have a truly agile cloud.

This implies that there needs to be a standard VM runtime environment in use between Microsoft, Xen, and VMware based clouds. We know this is not going to happen anytime soon. The best we can hope for is that the Distributed Management Task Force (DMTF) and its Open Virtualization Format (OVF) Specification can be used to import and export VMs between disparate clouds. However, for this to happen, each vendor (Microsoft, Xen, VMware, Red Hat) needs to support OVF imports and exports without changing any of the data and VM metadata as it is imported or exported. VMware makes heavy use of OVF in their products while the other vendors are lagging behind. However, even so, VMware’s use of OVF for vApps is pushing the boundaries as Microsoft, Xen, and Red Hat do not have similar functionality within their products.

The big three VMware, Microsoft, and Citrix are on the DMTF board and Red Hat is a member of the Leadership. Hopefully, these four can come together with respect to cloud and develop a truly agile and interoperable environment that will work over a set of islands.

Share this Article:

Share Button
Edward Haletky (368 Posts)

Edward L. Haletky, aka Texiwill, is the author of VMware vSphere(TM) and Virtual Infrastructure Security: Securing the Virtual Environment as well as VMware ESX and ESXi in the Enterprise: Planning Deployment of Virtualization Servers, 2nd Edition. Edward owns AstroArch Consulting, Inc., providing virtualization, security, network consulting and development and The Virtualization Practice where he is also an Analyst. Edward is the Moderator and Host of the Virtualization Security Podcast as well as a guru and moderator for the VMware Communities Forums, providing answers to security and configuration questions. Edward is working on new books on Virtualization.

[All Papers/Publications...]

Connect with Edward Haletky:


p5rn7vb

Related Posts:

Leave a Reply

Your email address will not be published. Required fields are marked *


× 3 = nine