Comments on: Long Distance vMotion… heading to the vCloud http://vinf.net/2009/06/30/long-distance-vmotion-heading-to-the-vcloud/ My ramblings on the stuff that holds it all together Tue, 14 Feb 2012 12:29:05 +0000 hourly 1 http://wordpress.com/ By: Geographically dispersed cluster design | Arnim van Lieshout http://vinf.net/2009/06/30/long-distance-vmotion-heading-to-the-vcloud/#comment-1248 Mon, 16 Nov 2009 11:46:43 +0000 http://vinf.wordpress.com/2009/06/30/long-distance-vmotion-heading-to-the-vcloud/#comment-1248 [...] http://www.virtuallifestyle.nl/2009/09/vmworld-09-long-distance-vmotion-ta3105/ http://vinf.net/2009/06/30/long-distance-vmotion-heading-to-the-vcloud/ http://www.yellow-bricks.com/2009/09/21/long-distance-vmotion/ [...]

]]> By: WAN VMotion - A Step Closer to a Private Cloud? | The SLOG - SimonLong/Blog http://vinf.net/2009/06/30/long-distance-vmotion-heading-to-the-vcloud/#comment-925 Tue, 30 Jun 2009 09:55:00 +0000 http://vinf.wordpress.com/2009/06/30/long-distance-vmotion-heading-to-the-vcloud/#comment-925 [...] from vinf.net recently commented on his blog post about this proof of concept, tells us; Before I had my GbE switch at home I could easily do [...]

]]> By: vinf http://vinf.net/2009/06/30/long-distance-vmotion-heading-to-the-vcloud/#comment-924 Tue, 30 Jun 2009 08:36:03 +0000 http://vinf.wordpress.com/2009/06/30/long-distance-vmotion-heading-to-the-vcloud/#comment-924 I wonder if it will be latency or bandwidth that kills it?

Before I had my GbE switch at home I could easily do vMotions between hosts over an old 100mb hub, 100mb WAN connections (at least within the UK) aren’t beyond the realms of affordability for many organisations and 1Gb LES links also not too bad within the M25

The article doesn’t make this clear but I assume it’s still using a single piece of storage which is accessible from both DC’s, if VMware could make long distance vMotion understand array mirroring (or via 3rd party storage virtualization like datacore etc.) then the actual bandwidth between the two DC’s would be slightly less of a concern as ongoing execution will happen from DC-local storage and you rely on the array to ship the data to the secondary node’s storage. this would need some kind of checkpoint/control mechanism which ESX controls – as long as you can sync the LUN/vmfs delta’s between the two sites at the storage level with an acceptable latency for vMotion (and I’ve made this work over a 100mb hub with ESX 3.5 before now)

This is similar to how EMC implemented long distance/stretched failover clusters with Exchange 2003/SQL – there was a plug-in product that handled the checkpointing of replication and reporting to the host.

Maybe not within the current release of the product, unless the new vSphere SATP/PSP modules can implement this? – VMware – feature request? ;)

]]> By: Simon Long http://vinf.net/2009/06/30/long-distance-vmotion-heading-to-the-vcloud/#comment-923 Tue, 30 Jun 2009 08:13:13 +0000 http://vinf.wordpress.com/2009/06/30/long-distance-vmotion-heading-to-the-vcloud/#comment-923 Hi Simon, i have also just been reading this article. I think probably the biggest issue companies will have is available bandwith between DC’s.

I would be interested to see what the minimum bandwidth is needed to run a successful vMotion.

Simon

]]>