What an amazing week in Washington D.C. for Nutanix .NEXT 2017! Although much of the focus was around our partnership with Google, Nutanix Calm automation and Xi cloud services, we also made plenty of really cool advancements in core Acropolis and AHV features and functions that will be available in a forthcoming release. All off this is part of our vision and journey to the Enterprise Cloud that starts with having a Cloud OS that provides a single fabric to bridge the gap between public and private clouds. One OS, One Click. If you were unable to join us in DC, here is a link to our Chief Product and Development Officer, Sunil Potti’s keynote that ties it all together.
In general the theme of the next planned release will be “enterprise cloud,” the features and functions that will be included in the releases will serve as connection points to Calm and Xi and will round out all the fundamental features required to operate and manage a hybrid enterprise cloud. The title says “summary” but this blog is packed full of features and enhancements to make things a little easier to digest it is organized into the following categories: AOS Core Datapath, Prism, AHV, Networking, and Tooling & Technologies.
Core Data Path (CDP)m
Core data path encompasses all the foundational hyper-converged infrastructure (HCI) functionality found in Acropolis Operating System (AOS) software, including storage services, availability, and data protection. Below you will find descriptions of a few of the new CDP features currently under development that were covered during the keynotes.
Near Sync Replication
AOS has long included the capability for synchronous and asynchronous replication as part of base data protection features. The existing sync-rep features provide for a near 0 RPO but are limited by the physics of bandwidth, latency, and distance. Async-rep does not have those limitations but was limited to a 60 minute RPO.
At the show we announced a major technology improvement called near-sync, which enables 1 minute RPO configurations without restrictions on latency or distance. For near-sync, AOS is leveraging a new Light Weight Snapshot (LWS) technology. LWS reduces overhead in managing metadata and increases storage performance by reducing the high number of storage IO operations caused by long snapshot chains.
Check out this blog post by Dwayne Lessner (@dlink7) for some additional detail.
Software Data at Rest Encryption (DaRE)
Nutanix has had support for hardware based encryption based on SED drives since our 4.1 release in 2014. What we will be introducing is the ability to deliver a similar level of security via a software solution without the additional expense of the SED drives. The initial release of the new solution will still require the use of a third party key management server (KMS), but native key management support will follow soon after. This solution will still be FIPS 140-2 certified, meeting the most stringent encryption regulations.
1- and 2-Node Remote, Branch and Edge Configurations
Remote office configurations can vary greatly based on workloads and business cases being solved. Currently, a Nutanix solution has the minimum requirement of 3 physical node (servers) to create a standard cluster. This is mostly due to how the local data protection and metadata functions are delivered. Requiring three nodes caused a situation where though there was a desire for an HCI solution with the elegance of Nutanix management and data protection features it could be difficult to justify three nodes for the smaller remote or edge workloads. The hardware and resource capacity would far outweigh the need.
At .NEXT we announced a new platform with support for 1- or 2-node deployments that is better suited to the capacity and availability requirements in these remote locations. The added benefit is that the ROBO and general datacenter deployed Nutanix clusters will all be centrally managed through Prism and leverage a common set of replication and data protection features.
Prism and Prism Pro
An overarching theme of One click, One OS was heard at .NEXT and our Prism Central (PC) management is the focus of that strategy. PC provides multi-cluster management and is the aggregation point for performance and usage metrics. This makes it the ideal location for our existing and future analytics based on our patent pending X-fit machine learning technology that are included as part of the Prism Pro feature set. In this section, I’ll cover a number of important enhancements that will be released in Prism.
Prism Central 1-Click Deploy
With the importance of PC, we had to make it easier for our customers to deploy. Existing customers already have Prism Element management available on every Nutanix node as a component of our controller VM (CVM) and our next release will include 1-click deployment for PC. The new PC deployment model will allow for a distributed and scalable PC that mimics how the Nutanix scales the virtualization and storage nodes. PC can be deployed as a single VM or for higher availability and increased scale, a 3 node clustered PC instance can be deployed automatically. This is a stark contrast to other virtualization management platforms that are lacking in availability features and complex to scale.
Prism Pro Scheduled Reporting
Communication is the key to any successful enterprise. In the IT world there are many stakeholders that would benefit from operational insights from the virtualization platform. Our customers know that Prism and Prism Central provide very rich monitoring and analytics, but that information is only available via the GUI. The next release will leverage existing custom dashboards and other reports combined with the ability to schedule the creation and delivery of email based reports based on that content.
Prism Pro VM Right Sizing
Entire products have been created around solving the problems of inefficient resource allocation and usage. Who hasn’t been in the situation where an application owner insists on a large resource allocation and only later to find that the VM never used a fraction of the RAM or CPU requested or spending time hunting for those unused or zombie VMs that were decommissioned but never really deleted? This is an area where our Xfit analytics and tightly integrated HCI platform really shine.
With such granular visibility across both virtualization and storage, it will be possible to create highly accurate VM sizing recommendations. Prism Pro will report on those zombie VMs and identify both over provisioned or resource constrained VMs and provide a recommended ideal allocation based on historical and observed behavioral characteristics. This gives the operations teams and application owners a high level of confidence in the applications getting peak performance while also optimizing VM density which in the end is a cost savings.
Prism Pro Dynamic Alerting / Anomaly Detection
Traditional monitoring based on thresholds and baselines are a working solution but there are many situations it’s not enough. Consider a sudden jump or shift in a performance metric that is still below a configured threshold, no alarm would be triggered. This could be a symptom of an impending problem that an operator would prefer to address before it becomes an emergency. Its no surprise that Nutanix is introducing dynamic threshold based alerting. Operators can now be alerted when key performance indicators (KPI) deviate from the norm. This takes the guesswork out of alert/alarm configuration and filtering.
Prism Pro Scheduled Upgrades
Some features and functions are cooler than others. But even the most mundane tasks can be huge consumers of valuable staff resources. Prism Central has always been our solution for multi-cluster monitoring. In our 5.x release it also became a central location for VM management for both AHV and ESXi based VM. Each of these improvements has made multi-cluster Nutanix management simpler and more efficient. Continuing on that course we will expand on the existing ability to kick of individual cluster upgrades with schedules. By adding schedules the operators can pre-stage upgrades during planned maintenance across 1 or more clusters. The enhancement will allow granular configuration of the process that includes choice of serial vs parallel upgrades and the ability to pause further actions should the upgrades cross out of the planned window. This automation frees up operators to spend time on higher value tasks vs just babysitting an upgrade progress bar.
AHV has seen a massive number of enhancements and improvements over the last few releases. This is apparent to our customers (over 23% of nodes shipped last quarter, based on a rolling four quarter average, are running AHV) that are using it, but AHV is not just the open source KVM hypervisor with some fancy dressing. Our goal from the beginning was two fold: provide the same level of consumer grade management for virtualization that has been delighting our customers that use Prism for our HCI platform management and leverage AHV as a deeper integration point for increased visibility and performance.
During the keynote, Sunil spoke about traditional hypervisor platforms being designed for a 3-tier SAN world. What isn’t as obvious is that as we move into new storage and memory technologies, the workloads need to be as close to the data as possible. Technologies like NVMe and RDMA have higher bandwidth and lower latencies than a network can provide. Though our current platform provides superior performance using the I/O data paths inherited from its KVM roots, it was obvious that we would need to make improvements for our pending support for NVMe and RDMA based hardware. From this need, AHV Turbo (internally called FRODO) was born. AHV Turbo will replace the standard KVM single queue/thread data path that was designed to generically support multiple virtual storage options with a purpose built high performance IO interface. Much like other enhancements in the Nutanix world, using AHV Turbo is non-disruptive for the operator. The only requirement is to upgrade the AOS and AHV software and instantly have a higher performing virtualization layer. If you’re curious about the details, check out this blog post for more.
If you blinked, you may have missed this important enhancement. AHV has already proven to be the hypervisor of choice for our customers deploying Citrix XenApp and XenDesktop VDI solutions, but for some use cases AHV was lacking a key function: vGPU support. We can now say that support is officially announced and customers will be able to leverage nVidia Tesla M10 and M60 GPU cards to enable higher desktop density, better graphic performance for VDI, and support high end graphics use cases like CAD when deploying Citrix VDI on Nutanix with the AHV hypervisor.
NX-9030 (RDMA & NVMe)
With the enhancements from AHV Turbo, there had to be some cool new hardware to take advantage of. We announced our next new platform focused on tier 0/tier 1 applications, the NX-9030, that will provide support for 40Gb Ethernet, NVMe flash storage and RDMA. RDMA has particular benefit to our architecture with a CVM located on each node. Without RDMA all CVM to CVM communication must traverse the virtualization stack and then go over a standard Ethernet network to a top-of-rack (ToR) switch. I mentioned earlier to take advantage of new storage technology like NVMe and the upcoming Intel 3D Xpoint new solutions would be required. We solved a portion of that with AHV Turbo, but by leveraging RDMA, CVM to CVM communications can bypass the hypervisor and copy data directly thus allowing NDFS optimal access to the new high speed storage.
Nutanix AHV started out with the standard virtual network constructs and some really cool IP address management features (IPAM). Continuing on our convergence journey we are now expanding our software defined networking (SDN) features. What you may not know is that we use Open vSwitch (OVS) and Open Virtual Network (OVN) as our networking core. OVS and OVN share many of the same properties and features as VMware’s NSX, both having the same creator, Martin Casado. What this means for our customers is that we have a common set of primitive functions that we can both expand on and apply our Prism management 1-click philosophy to. At .NEXT we introduced the three following networking features. This is only the beginning of our networking expansion.
We like to talk about all the simplification benefits of HCI with regard to compute and storage, but we all know that the glue that puts it all together is the network. HCI greatly simplifies networking, as it removes the need for separate storage networks and with Nutanix even much of the network overhead because of our data locality feature. That said, there is still plenty more we can do to make the network easier to manage. We introduced a set of APIs in our AOS 5.0 release that allows 3rd party vendors do subscribe to VM provisioning and network change events. This allows the network to react to changes on the virtual machine. For example, updating a load balancing policy or altering security rules.
We have a number of our ecosystem vendors in our Nutanix Ready program that have plans to implement this functionality. To list a few: Arista, Big Switch, Brocade, Citrix, Cumulus, F5, Lenovo, Mellanox, Palo Alto Networks, Plexxi, and vArmour. Look for more announcements and specifics soon.
Service Insertion for Network Function Virtualization (NFV)
I spoke of automation above, but a key part of that automation are the functions and services provided by those ecosystem vendors. A big part of being “software defined” is the ability to surgically insert and remove services for various workloads. For the class of services that operate on network traffic, Nutanix provides a set of APIs that allows programmatic insertion of our ecosystem partner network services and set policy for how they are chained together.
This gets really cool when tied to the automation functions above. Consider an application blueprint in Nutanix Calm that specifies not just the policies, but the 3rd party service to use for enforcement. That could be security from a Palo Alto firewall or load balancing rules for a Citrix Netscaler. From the application owner’s perspective, they just create the VMs, all the other network configuration and connection happens like magic!
Traditional perimeter firewalls are great for edge defence, but those that practice security talk about “defence in depth”. This is the concept that a true security policy needs a multi-layer strategy. A common saying was it doesn’t matter if you lock the doors if all the windows are wide open. When thinking about granular network controls at the server or VM level, many have defaulted to a host based firewall. The problem with this approach is that it becomes a management challenge at scale. Even if the server OS provides the native functionality vs installing a software agent, there is still a huge overhead of policy management.
Our solution will be two fold, we leverage the fine grain controls provide by the OVS networking and couple it with a “”1-click” policy approach that separates the network attributes from the VMs and Applications. Policy makers can group VMs and Applications into categories that policies are applied to. For example, a fence can be drawn around Test/Dev and Production and then more specific policies for databases are layered in. Removing the need to know or track network details like IP addresses simplifies policy writing and enables mobility without the worry that security may be impacted.
Tooling & Technologies
Beyond the core products Nutanix offers, we are also providing and developing tools to help customers get a quicker time to value, deploy using our best practices, and have better visibility into performance. At the conference we announced two new tools to help in this area: Xtract for migration and data transformation, and X-ray to allow benchmarking of not just Nutanix but other HCI offerings as well.
With all of the amazing features and enhancements discussed in Nutanix Enterprise Cloud, it’s not surprising that we are frequently asked by customers “How can I migrate my VMs over?”. Xtract is one of those methods, comprising of two offerings i: Xtract for VMs, currently under development, and Xtract for DBs, already available for download at the Nutanix Portal Site. The names are fairly self explanatory with the VM product enabling project based bulk VM migration in a ‘lift-and-shift’ fashion, from VMware ESXi to AHV with minimal service disruption. Xtract for DBs’ transformational capabilities work at the application-layer, discovering and assessing Microsoft SQL Servers regardless of source platform (physical, virtual, and even public cloud based). Applying industry best practises to their configuration and performance data, a design template is created, whereby new SQL Server VMs are instantiated on Nutanix Enterprise Cloud, and DB data is migrated into these new VMs. Both VM and DB capabilities enable full cut-over control with licenses included for all PRISM users, saving businesses the cost and manual effort of migrating workloads by other means. Read further in this blog post.
The complexity of SAN storage led to a huge focus on performance and benchmarking. That focus created an entire body of testing tools with various approaches. To this day there is still much debate on synthetic testing vs real world observations. What is not in question is that HCI has changed the game with existing tools not being suited for an HCI test. To address this, Nutanix created X-Ray which is designed to use real world scenarios for both performance and availability (e.g. upgrades, hardware failure, etc) to create a benchmark or balanced comparison between platforms. You can read more here or download it and check it out for yourself.
I’m hesitant to end with “Well that’s it.” as this blog covers over 15 new products and features that are part of our upcoming releases and we didn’t even get into Calm and Xi! Thanks for sticking with me to the end and I hope that we have sparked your interest in what’s to come. Looking forward to seeing all of you at .NEXT Europe in Nice in September or back in the states .NEXT year in New Orleans! Stay tuned to our twitter feed (@nutanix), website and customer forums (my.nutanix.com) for more details and updates!
Forward Looking Statements
This blog includes forward-looking statements, including but not limited to statements concerning our plans and expectations relating to product features and technology that are under development or in process and capabilities of such product features and technology and our plans to introduce product features in future releases. These forward-looking statements are not historical facts, and instead are based on our current expectations, estimates, opinions and beliefs. The accuracy of such forward-looking statements depends upon future events, and involves risks, uncertainties and other factors beyond our control that may cause these statements to be inaccurate and cause our actual results, performance or achievements to differ materially and adversely from those anticipated or implied by such statements, including, among others: failure to develop, or unexpected difficulties or delays in developing, new product features or technology on a timely or cost-effective basis; delays in or lack of customer or market acceptance of our new product features or technology; the introduction, or acceleration of adoption of, competing solutions, including public cloud infrastructure; a shift in industry or competitive dynamics or customer demand; and other risks detailed in our Form 10-Q for the fiscal quarter ended April 30, 2017, filed with the Securities and Exchange Commission. These forward-looking statements speak only as of the date of this presentation and, except as required by law, we assume no obligation to update forward-looking statements to reflect actual results or subsequent events or circumstances.
Disclaimer: This blog may contain links to external websites that are not part of Nutanix.com. Nutanix does not control these sites and disclaims all responsibility for the content or accuracy of any external site. Our decision to link to an external site should not be considered an endorsement of any content on such site.
© 2017 Nutanix, Inc. All rights reserved. Nutanix, the Enterprise Cloud Platform, Xi, Nutanix Calm, the Nutanix logo and other Nutanix products mentioned are registered trademarks or trademarks of Nutanix, Inc. in the United States and other countries. All other brand and product names mentioned herein are for identification purposes only and may be the trademarks of their respective holder(s).