Date:   Thu, 5 Aug 2021 20:34:23 +0200
From:   Greg Kroah-Hartman <gregkh@...uxfoundation.org>
To:     Long Li <longli@...rosoft.com>
Cc:     Bart Van Assche <bvanassche@....org>,
        "longli@...uxonhyperv.com" <longli@...uxonhyperv.com>,
        "linux-block@...r.kernel.org" <linux-block@...r.kernel.org>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        "linux-hyperv@...r.kernel.org" <linux-hyperv@...r.kernel.org>,
        Jonathan Corbet <corbet@....net>,
        KY Srinivasan <kys@...rosoft.com>,
        Haiyang Zhang <haiyangz@...rosoft.com>,
        Stephen Hemminger <sthemmin@...rosoft.com>,
        Wei Liu <wei.liu@...nel.org>, Dexuan Cui <decui@...rosoft.com>,
        Bjorn Andersson <bjorn.andersson@...aro.org>,
        Hans de Goede <hdegoede@...hat.com>,
        "Williams, Dan J" <dan.j.williams@...el.com>,
        Maximilian Luz <luzmaximilian@...il.com>,
        Mike Rapoport <rppt@...nel.org>,
        Ben Widawsky <ben.widawsky@...el.com>,
        Jiri Slaby <jirislaby@...nel.org>,
        Andra Paraschiv <andraprs@...zon.com>,
        Siddharth Gupta <sidgup@...eaurora.org>,
        Hannes Reinecke <hare@...e.de>
Subject: Re: [Patch v5 0/3] Introduce a driver to support host accelerated
 access to Microsoft Azure Blob for Azure VM

On Thu, Aug 05, 2021 at 06:24:57PM +0000, Long Li wrote:
> > Subject: Re: [Patch v5 0/3] Introduce a driver to support host accelerated
> > access to Microsoft Azure Blob for Azure VM
> > 
> > On 8/5/21 12:00 AM, longli@...uxonhyperv.com wrote:
> > > From: Long Li <longli@...rosoft.com>
> > >
> > > Azure Blob storage [1] is Microsoft's object storage solution for the
> > > cloud. Users or client applications can access objects in Blob storage
> > > via HTTP, from anywhere in the world. Objects in Blob storage are
> > > accessible via the Azure Storage REST API, Azure PowerShell, Azure
> > > CLI, or an Azure Storage client library. The Blob storage interface is
> > > not designed to be a POSIX compliant interface.
> > >
> > > Problem: When a client accesses Blob storage via HTTP, it must go
> > > through the Blob storage boundary of Azure and get to the storage
> > > server through multiple servers. This is also true for an Azure VM.
> > >
> > > Solution: For an Azure VM, the Blob storage access can be accelerated
> > > by having Azure host execute the Blob storage requests to the backend
> > > storage server directly.
> > >
> > > This driver implements a VSC (Virtual Service Client) for accelerating
> > > Blob storage access for an Azure VM by communicating with a VSP
> > > (Virtual Service
> > > Provider) on the Azure host. Instead of using HTTP to access the Blob
> > > storage, an Azure VM passes the Blob storage request to the VSP on the
> > > Azure host. The Azure host uses its native network to perform Blob
> > > storage requests to the backend server directly.
> > >
> > > This driver doesn't implement Blob storage APIs. It acts as a fast
> > > channel to pass user-mode Blob storage requests to the Azure host. The
> > > user-mode program using this driver implements Blob storage APIs and
> > > packages the Blob storage request as structured data to VSC. The
> > > request data is modeled as three user provided buffers (request,
> > > response and data buffers), that are patterned on the HTTP model used
> > > by existing Azure Blob clients. The VSC passes those buffers to VSP for Blob
> > storage requests.
> > >
> > > The driver optimizes Blob storage access for an Azure VM in two ways:
> > >
> > > 1. The Blob storage requests are performed by the Azure host to the
> > > Azure Blob backend storage server directly.
> > >
> > > 2. It allows the Azure host to use transport technologies (e.g. RDMA)
> > > available to the Azure host but not available to the VM, to reach to
> > > Azure Blob backend servers.
> > >
> > > Test results using this driver for an Azure VM:
> > > 100 Blob clients running on an Azure VM, each reading 100GB Block Blobs.
> > > (10 TB total read data)
> > > With REST API over HTTP: 94.4 mins
> > > Using this driver: 72.5 mins
> > > Performance (measured in throughput) gain: 30%.
> > >
> > > [1]
> > >
> > https://nam06.safelinks.protection.outlook.com/?url=https%3A%2F%2Fdocs
> > > .microsoft.com%2Fen-us%2Fazure%2Fstorage%2Fblobs%2Fstorage-blobs-
> > intro
> > >
> > duction&amp;data=04%7C01%7Clongli%40microsoft.com%7C6ba60a78f4e74
> > aeb0b
> > >
> > b108d95833bf53%7C72f988bf86f141af91ab2d7cd011db47%7C1%7C0%7C6376
> > 378015
> > >
> > 92577579%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoi
> > V2luMzIiL
> > >
> > CJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C1000&amp;sdata=ab5Zl2cQdmUhdT3l
> > SotDwMl
> > > DQuE0JaY%2B1REPQ0%2FjXa4%3D&amp;reserved=0
> > 
> > Is the ioctl interface the only user space interface provided by this kernel
> > driver? If so, why has this code been implemented as a kernel driver instead
> > of e.g. a user space library that uses vfio to interact with a PCIe device? As an
> > example, Qemu supports many different virtio device types.
> 
> The Hyper-V presents one such device for the whole VM. This device is used by all processes on the VM. (The test benchmark used 100 processes)
> 
> Hyper-V doesn't support creating one device for each process. We cannot use VFIO in this model.

I still think this "model" is totally broken and wrong overall.  Again,
you are creating a custom "block" layer with a character device, forcing
all userspace programs to use a custom library (where is it at?) just to
get their data.

There's a reason the POSIX model is there, why are you all ignoring it?

greg k-h

Powered by blists - more mailing lists