Welcome!

Java IoT Authors: Liz McMillan, Yeshim Deniz, Zakia Bouachraoui, Elizabeth White, Pat Romanski

Related Topics: @CloudExpo, Java IoT, Containers Expo Blog

@CloudExpo: Blog Post

Predictable Performance & Granular Control By @KellyMurphyGS | @CloudExpo [#Cloud]

In a highly virtualized environment, delivering predictable performance to VMs from a shared elastic resource is critical

Design Goal 2: Predictable Performance and Granular Control per VM

Traditional storage was designed to present a LUN to a server. This worked well until that server became a hypervisor running 10, 20 or 30 virtual servers, all pointing to the same LUN. Adding to the complexity, all hypervisors need to point to the same LUN to enable VM mobility or high availability. Architecturally, this creates a many-to-one relationship that makes optimization impossible. Each VM has a different I/O pattern. When they all mix together, the aggregate I/O becomes highly random, creating performance issues with no way to isolate problems or control I/O per VM.

Isolate and Optimize I/O per VM
The core component of the Gridstore architecture is the vController that operates in the host kernel and appears as a local SCSI device. From this privileged position, it isolates I/O from each of the VMs into independent swim lanes or channels from the hypervisor across the network and onto storage nodes in the Grid. This re-creates the1:1 relationship between VM and storage. Once I/O is isolated, each channel can be independently optimized according to what the VM is trying to achieve. Optimizations occur dynamically optimizing the I/O pattern to the demands of the VM.

System Center
All-Flash Hyper-Converged Infrastructure
Predictable Performance per VM
With the I/O isolated into channels per VM, the vController can now overlay QoS policies or set specific IOPS minimum/maximum per VM to give predictable performance per VM. Upon provisioning VMs through Virtual Machine Manager (VMM) and classifying them as a platinum service, Gridstore will deliver a platinum storage service. Furthermore, administrators can dial in a specific amount of IOPS to a given VM to ensure that it allocates a minimum amount of I/O to enable it to run optimally. Equally, lower-grade classes can be created or limits on IOPS can be dialed in to ensure that no VM consumes more than it requires. This is beneficial for service providers who want to offer different levels of service to customers, since it ensures they can deliver what they promise, resulting in predictable environments that run as desired for all workloads.

End-to-End QoS Architecture
Due to the unpredictable performance in virtual environments, QoS has become a hot feature. However not all QoS is the same. Three models for QoS are:

  1. Hypervisor based
  2. Array based
  3. End-to-end

A. Hypervisor- (Host-) Based QoS
Hypervisors create QoS in the hypervisor by throttling I/O. This is effective to ensure that one or more VMs do not consume all the IOPS, the noisy neighbor problem. This model, however, only operates in a single hypervisor host. It does not provide a global view (i.e., storage resources being consumed by other hosts, like a database on a physical server). Furthermore, a hypervisor-based model is not in control of the actual storage resource, making it impossible to allocate more resources to a particular VM or physical host to meet a minimum IOPS reserve.

B. Array-Based QoS
This is the opposite of the hypervisor model. In this configuration a storage array presents multiple LUNs to the hypervisor. Each LUN has a QoS priority attached to it, and I/O is more or less "allocated" to these LUNs based on the prioritization. Problems with this model include:

  • Being LUN based, it is not granular to the VM level
  • The I/O blender is still blending, which means I/O that passes through a LUN is blended together from multiple VMs, resulting in no per-VM granularity. With this, hundreds of VMs on that LUN compete for resources. When a VM on the highest priority LUN runs wild and consumes all resources, every other VM is impacted. According to the QoS policy, it did its job. It served the highest-priority LUN and starved the other LUNs. However, it does not know that all this I/O is coming from a single VM at the expense of others. It does not know that they are on the same highest-priority LUN
  • An array-based QoS model does not have any ability to control the flow from the hypervisor end. Rather, its domain of control is within the array.

C. End-to-End Per-VM I/O Control
In a highly virtualized environment, delivering predictable performance to VMs from a shared elastic resource is critical. Gridstore's vController technology isolates I/O from each VM at the source and enables end-to- end per-VM I/O control through the entire storage stack. By controlling both the resource consumption (VM hosts) and the resource allocation (storage resources from Grid nodes), Gridstore can precisely monitor demand as well as control the allocation of resources across the hypervisor hosts and storage resources. End-to-end VM I/O control enables the ability to dial in specific I/O per VM or class of VMs, thus ensuring SLAs are met. The result: complete control of resource allocation and utilization.

The post Design Goal 2: Predictable Performance and Granular Control per VM appeared first on Gridstore.

More Stories By Kelly Murphy

As a serial entrepreneur with a track record of bringing disruptive technologies to market, Kelly Murphy brings 15 years CEO experience with disruptive venture backed software companies. In 1998, almost a decade before the cloud became popular, Murphy founded Marrakech, the first software company that offered on-demand procurement and supply chain systems to over 30,000 trading partners including some of the world’s largest retailers, consumer food producers, packaging companies and utilities. After selling Marrakech in 2007, he turned his sights onto what was his largest obstacle in growing his previous business — storage.

In 2009, Murphy founded Gridstore — a pioneer of software-defined storage that is set to disrupt the traditional storage industry. Currently, he serves on Gridstore’s Board of Directors and is also the Chief Technology Officer. Originally from Canada, Murphy obtained his BS in Computer Science from Michigan Technological University, played Division I hockey and was the seventh pick of the New York Islanders in the 1984 entry draft.

Comments (0)

Share your thoughts on this story.

Add your comment
You must be signed in to add a comment. Sign-in | Register

In accordance with our Comment Policy, we encourage comments that are on topic, relevant and to-the-point. We will remove comments that include profanity, personal attacks, racial slurs, threats of violence, or other inappropriate material that violates our Terms and Conditions, and will block users who make repeated violations. We ask all readers to expect diversity of opinion and to treat one another with dignity and respect.


IoT & Smart Cities Stories
CloudEXPO New York 2018, colocated with DXWorldEXPO New York 2018 will be held November 11-13, 2018, in New York City and will bring together Cloud Computing, FinTech and Blockchain, Digital Transformation, Big Data, Internet of Things, DevOps, AI, Machine Learning and WebRTC to one location.
Bill Schmarzo, Tech Chair of "Big Data | Analytics" of upcoming CloudEXPO | DXWorldEXPO New York (November 12-13, 2018, New York City) today announced the outline and schedule of the track. "The track has been designed in experience/degree order," said Schmarzo. "So, that folks who attend the entire track can leave the conference with some of the skills necessary to get their work done when they get back to their offices. It actually ties back to some work that I'm doing at the University of San...
Andrew Keys is Co-Founder of ConsenSys Enterprise. He comes to ConsenSys Enterprise with capital markets, technology and entrepreneurial experience. Previously, he worked for UBS investment bank in equities analysis. Later, he was responsible for the creation and distribution of life settlement products to hedge funds and investment banks. After, he co-founded a revenue cycle management company where he learned about Bitcoin and eventually Ethereal. Andrew's role at ConsenSys Enterprise is a mul...
IoT is rapidly becoming mainstream as more and more investments are made into the platforms and technology. As this movement continues to expand and gain momentum it creates a massive wall of noise that can be difficult to sift through. Unfortunately, this inevitably makes IoT less approachable for people to get started with and can hamper efforts to integrate this key technology into your own portfolio. There are so many connected products already in place today with many hundreds more on the h...
DXWorldEXPO | CloudEXPO are the world's most influential, independent events where Cloud Computing was coined and where technology buyers and vendors meet to experience and discuss the big picture of Digital Transformation and all of the strategies, tactics, and tools they need to realize their goals. Sponsors of DXWorldEXPO | CloudEXPO benefit from unmatched branding, profile building and lead generation opportunities.
DXWorldEXPO LLC announced today that Telecom Reseller has been named "Media Sponsor" of CloudEXPO | DXWorldEXPO 2018 New York, which will take place on November 11-13, 2018 in New York City, NY. Telecom Reseller reports on Unified Communications, UCaaS, BPaaS for enterprise and SMBs. They report extensively on both customer premises based solutions such as IP-PBX as well as cloud based and hosted platforms.
In his keynote at 19th Cloud Expo, Sheng Liang, co-founder and CEO of Rancher Labs, discussed the technological advances and new business opportunities created by the rapid adoption of containers. With the success of Amazon Web Services (AWS) and various open source technologies used to build private clouds, cloud computing has become an essential component of IT strategy. However, users continue to face challenges in implementing clouds, as older technologies evolve and newer ones like Docker c...
The best way to leverage your Cloud Expo presence as a sponsor and exhibitor is to plan your news announcements around our events. The press covering Cloud Expo and @ThingsExpo will have access to these releases and will amplify your news announcements. More than two dozen Cloud companies either set deals at our shows or have announced their mergers and acquisitions at Cloud Expo. Product announcements during our show provide your company with the most reach through our targeted audiences.
To Really Work for Enterprises, MultiCloud Adoption Requires Far Better and Inclusive Cloud Monitoring and Cost Management … But How? Overwhelmingly, even as enterprises have adopted cloud computing and are expanding to multi-cloud computing, IT leaders remain concerned about how to monitor, manage and control costs across hybrid and multi-cloud deployments. It’s clear that traditional IT monitoring and management approaches, designed after all for on-premises data centers, are falling short in ...
The deluge of IoT sensor data collected from connected devices and the powerful AI required to make that data actionable are giving rise to a hybrid ecosystem in which cloud, on-prem and edge processes become interweaved. Attendees will learn how emerging composable infrastructure solutions deliver the adaptive architecture needed to manage this new data reality. Machine learning algorithms can better anticipate data storms and automate resources to support surges, including fully scalable GPU-c...