At VMworld, various cool new technologies were previewed. In this series of articles, I will write about some of those previewed technologies. Unfortunately, I can’t cover them all as there are simply too many. This article is about enhancements in the business continuity/disaster recovery space. There were 2 sessions where futures were discussed, namely HCI2894BU and HBI3109BU. Please note that this is a brief summary of those sessions, and these are discussing a Technical Preview, these features/products may never be released, and these previews do not represent a commitment of any kind, and this feature (or it’s functionality) is subject to change. Now let’s dive into it, what can you expect for disaster recovery in the future?
The first session I watched was HCI2894BU, this was all about Site Recovery Manager. I think the most interesting part is the future support for Virtual Volumes (vVols) for Site Recovery Manager. It may sound like something simple, but it isn’t. When the version of SRM ships that supports vVols keep in mind that your vVol capable storage system also needs to support it. At day 1 HPe Nimble, HPe 3PAR and Pure Storage will support it and Dell EMC and NetApp are actively working on support. The requirements are that the storage system needs to be vVols 2.0 compliant and support VASA 3.0. Before they dove into the vVols implementation, some history was shared and the current implementation. I found it interesting to know that SRM has over 25.000 customers and has protected more than 3.000.000 workloads over the last decade.
First of all, why is this important? Well, vVols already has support for replication, but the orchestration via SRM was not supported. Of course for customers who prefer to use SRM over a self-written solution support for vVols is crucial. vVols is a completely different beast than traditional storage. The communication from SRM does not happen via a so-called storage replication adaptor (SRA) for vVols but happens through the VASA (vSphere APIs for Storage Awareness) provider via SMS (Storage Management Service). VASA is also what vSphere uses to communicate with the vVols storage system. Which is great if you ask me, as it reduces complexity!
So how does this work? Well, when you start SRM and VASA is discovered it will send a discovery request and will receive all data needed to create a protection group and recovery plan:
- SRM gets all datastores from vCenter
- SRM gets all fault domains and storage containers
- SRM gets all replication groups and peers
- SRM gets all related VMs
- Gets all storage profiles of all replicated VMs
Of course, this will happen in both locations so a comprehensive recovery plan can be created. When a recovery needs to occur the steps are very similar to a traditional storage system. (See screenshot below.) What I think is important to know is that public API support for SRM and vVols will also be provided in the upcoming release, and in a subsequent PowerCLI release new PowerCLI cmdlets will also be delivered.
Next, a demo was shown of vVols and SRM. This was followed up by both HPe and Pure Storage explaining their vVols and SRM capabilities. That was it for HCI2894BU, let’s switch to HBI3109BU. There was just one announcement in this session, and I am going to limit it to that as I want to avoid duplication of content and explaining basic offerings like VMware Site Recovery (our DR as a Service solution). In short, the problem for some customers with our current DR as a Service offering is the fact that it requires you to have a VMware Cloud on AWS SDDC up and running. In some cases that SDDC may be idle and only used as a replication target. Which for Tier-1 workloads may be acceptable as they will require a low RPO and low RTO. But what about Tier 2 and Tier 3 workloads?
For Tier 2 and Tier 3 workloads having an SDDC up and running may simply too expensive or overkill. For that reason, VMware is investing in a Cost Optimized DR as a Service solution. With the cost-optimized solution, VMs would be replicated to lower-cost cloud storage, without the need for an active SDDC running. Note that this is not a feature that would normally replace the current DRaaS solution, it should complement the current solution. Tier 1 and business-critical apps should use the “Performance Optimized” option and Tier 2 and Tier 3 workloads can potentially use the Cost Optimized option in the near future. No dates around when (and if) this will be released were provided.
And that was it for VMware disaster recovery and business continuity reveals during VMworld 2019. Hope you find this series useful so far!