In this two-part article, we eschew the normal blue-sky 2010 technology predictions for something a bit more everyday that will affect your daily virtual life. Missed part one? Be sure to read it for predictions about a major vSphere point release, ESXi as the VMware standard and bigger, more capable VMs.
Crisis? What crisis?
I hope VMware will widen the scope of its storage options for its disaster recovery automation product, Site Recovery Manager (SRM), this year. At the moment SRM only works with storage array-based replication from the big storage vendors. So far, purely software-based replication of VMs from the likes of VizionCore and Veeam has been excluded from the SRM project.
I hope VMware removes this restriction to allow these companies to integrate with SRM. We might also see VMware develop its own replication engine into the ESX host to allow small and medium businesses (SMBs) to protect their investment in their virtual infrastructure. This is similar to how VMware has entered the SMB virtual machine (VM) backup market by releasing the VMware Data Recovery Appliance (vDR). The next logical step would be to offer those companies a built-in method to replicate their VMs to another location. The tricky part would be the DR integration for these SMBs. Implementing SRM is no light undertaking from a licensing perspective –- requiring a licensed vCenter and ESX hosts in the DR location. It would be nice to see companies such as VizionCore and Veeam offer an SRM-like product to the SMB market.
Backup and availability
I expect VMware to continue its commitment to the vDR backup appliance. Although the publically stated line from VMware is that it remains equally committed to the VMware Consolidated Backup (VCB) and vDR technologies, I suspect we will see more research and development on the vDR appliance with more changes and improvements to it. What I hope to see is that vDR appliance itself will be patched and fixed via the VMware Update Manager, rather than customers downloading a new appliance every time a fix is needed. Additionally, I suspect that VMware will have to standardize on a common operating system for its virtual appliances, as blue-chip companies that have a preexisting commitment to Linux react negatively to new distributions sneaking in through the back door in the form of virtual appliances.
I think that in the future VMware may want to enable the ability to run multiple versions of vDR in vCenter as a way of increasing the scalability of vDR. Currently, you are able to run multiple versions of vDR (there is no licensing restriction in that respect), but only one can be active and configured for vCenter, which is somewhat limiting. VMware's biggest backup challenge is trying to square VM backups with VMware's new Fault Tolerance (FT) solution.
Fault Tolerance, High Availability and snapshots
VMware FT is currently incompatible with VMware snapshots, and VMware snapshots form the backbone of any "hot backup" of a VM. This incompatibility currently limits the use VMware's VCB, vDR and third-party backup to tools from the likes of VizionCore, Veeam and PhD Technologies, and forces customers to revert back to installing backup agents into the guest operating system. Unfortunately, I don't see a solution for this issue coming until 2011.
On the subject of VMware High Availability (HA) and FT I would expect to see only slight improvements this year. What I was hoping to see is that VMware will provide quality assurance (QA) and support the use of VM clustering, such as Microsoft Cluster Server (MSCS) and Veritas clustering. It has been technically possible to use both VMware HA and third-party clustering technologies for some time, although VMware did not support this configuration. Ironically, I had expected to see a decline of interest in MSCS because of this -- but last year, I had many students and customers asking me about this configuration.
I think the reason for this renewed interest in clustering technologies inside a VM reflects two issues. Firstly, while VMware clustering is very good, it doesn't directly address the failure of a service within the guest operating system. Secondly, many physical clusters are nearing the end of their support cycle and folks have to ask themselves the question: should they stay on physical or should they be ported to a virtual world? If VMware changes its support stance that question becomes a much viable proposition.
Even so, I don't think the really interesting technological advances in VMware HA and FT will come until 2011, when we might just see VMware pushing the use of HA and FT across longer distances. Many people don't realize that the design of HA and FT really reflects a server availability technology rather than a site availability technology. I think VMware will want to push these technologies to cover great distances, just as they have begun to do with VMotion.
Increasingly as the virtualization story morphs into a cloud computing story, VMware will want to demonstrate that its platform can scale-up to this grand vision of providing computing power to large organizations. So in 2011, I would expect to see VMware HA's failover capacity, or the number of ESX hosts and VMs that can be protected, to increase accordingly. This is all a part of VMware's attempt to demonstrate that it has both scale-up and scale-out capabilities.
Because of these changes, I think there will be two more updates in 2010. I think we will see an Update 2 by the end of the second quarter and probably Update 3 by the end of 2010. The VMware's Vi3.5 platform saw a dizzying number of updates after its release; four in a single year. I think VMware will want to move to a less frequent bi-annual approach for future updates, releasing two updates per year. Many VMware shops found the update release strategy used by VMware in previous releases a burden. So I think that will put vSphere4.5 firmly into the 2011 time period, where these much bigger blue-sky thinking features will surface -- just in time for the annual VMworld conference. Historically, VMware has always had a major .5 release within its enterprise platform (2.0/2.5, 3.0/3.5), so I wouldn't be surprised to see this pattern repeat itself for vSphere 4. It generally comes 18 months or so after the main product launch.
As you can see, VMware wants to position itself as the de facto platform for scale-up and scale-out cloud computing, though perhaps at the risk of neglecting the SMB market, where historically VMware hasn't been strong player. Some might say this is quite a risky strategy, as no one yet seems to have a firm grip on what the "cloud" actually is or how it works. But it's a project that no one wants to be excluded from -– least of all the company that might just make it possible.
Mike Laverick (VCP) has been involved with the VMware community since 2003. Laverick is a VMware forum moderator and member of the London VMware User Group Steering Committee. Laverick is the owner and author of the virtualization website and blog RTFM Education, where he publishes free guides and utilities aimed at VMware ESX/VirtualCenter users, and has recently joined SearchVMware.com as an Editor at Large. In 2009, Laverick received the VMware vExpert award and helped found the Irish and Scottish VMware user groups. Laverick has had books published on VMware Virtual Infrastructure 3, VMware vSphere4 and VMware Site Recovery Manager.