Find Communities by: Category | Product

In my last blog, I discussed about the basics of Virtualization. In this blog, I will discuss the components, features and benefits of vSphere.

 

Today’s IT leaders have to balance the often-competing demands of delivering greater flexibility, availability and responsiveness to change, with the need to manage costs. To that end, trends in the datacenter point towards the growth in the use of virtualized servers exceeding that of physical servers. Here VMWare’s vSphere plays an important role. VMware vSphere is the brand name for VMware's suite of virtualization products. Before 2009, VMware vSphere was known as VMware Infrastructure.

VMware® vSphere® is an infrastructure virtualization suite that provides virtualization, management, resource optimization, application availability, and operational automation capabilities in an integrated package. vSphere virtualizes and aggregates the underlying physical hardware resources across multiple systems and provides pools of virtual resources to the datacenter. In addition, vSphere provides a set of distributed services that enable detailed, policy-driven resource allocation, high availability, and scalability of the entire virtual datacenter.

 

 

dd1.png

 

VMware vSphere, which is a necessary component of the vCloud Suite for cloud computing, includes

 

  • VMware ESXi -  abstracts processor, memory, storage, and other resources into multiple virtual machines (VMs).
  • VMware vCenter Server - central control point for data center services such as access control, performance monitoring and alarm management.
  • VMware vSphere Client - allows users to remotely connect to ESXi or vCenter Server from any Windows PC.
  • VMware vSphere Web Client - allows users to remotely connect to vCenter Server from a variety of Web browsers and operating systems (OSes).
  • VMware vSphere SDKs - provides interfaces for accessing vSphere components.
  • vSphere Virtual Machine File System (VMFS) - provides a high performance cluster file system for ESXi VMs.
  • vSphere Virtual SMP - allows a single virtual machine to use multiple physical processors at the same time.
  • vSphere vMotion - allows live migration for powered-on virtual machines in the same data center.
  • vSphere Storage vMotion -  allows virtual disks or configuration files to be moved to a new data store while a VM is running.
  • vSphere High Availability (HA) - allows virtual machines to be restarted on other available servers.
  • vSphere Distributed Resource Scheduler (DRS) - divides and balances computing capacity for VMs dynamically across collections of hardware resources.
  • vSphere Storage DRS - divides and balances storage capacity and I/O across collections of data stores dynamically.
  • vSphere Fault Tolerance - provides continuous availability.
  • vSphere Distributed Switch (VDS) -  allows VMs to maintain network configurations as the VMs migrate across multiple hosts.
  • Host Profiles - provides a way to create user-defined configuration policies.

 

 

VMWare vSphere undergoes periodic revisions and updates to add features, modifications to the application program interface (API) and changes to the ESXi Shell. VMware offers vSphere in several kits (Essentials and Essentials Plus) and editions (Standard, Enterprise and Enterprise Plus.)

 

 

Some Salient Features of  VMWARE vSphere

 

Below are some features of the vSphere which takes the product apart in the virtualization domain.

 

  • vSphere as a Software Define Data Center

 

The software defined datacenter is considered to be the foundation of cloud computing. The software defined datacenter deploys virtual datacenters with isolated computing, storage, networking and security resources faster than the traditional hardware based datacenter. vSphere is critical to the success of the software defined datacenter because it provides the hardware and networking abstraction and resource pooling necessary for the datacenter to deploy. Below diagram provides us some illustration:

dd2.png

 

 

 

  • vSphere fits into Cloud Computing

dd3.png

The consolidation and rapid virtual machine provisioning results in a high-level of server use and reuse that enables an effective use of capital equipment. When coupled with technologies such as virtual machine migration, high availability, integrated data protection and centralized management it is easy to see how vSphere 5 is the clear choice for a cloud computing platform.

 


  • Offers protection at every level with the allied tools of VMWare

dd4.png

At every level of the datacenter, from individual components all the way up to the entire site, VMware® vSphere® 5.x provides protection against both planned and unplanned downtime. All these features combine to provide greater availability to all supported operating systems and applications.Many methods ensure highly availability in a virtualized environment. vSphere 5.x uses technologies like the following to ensure that virtual machines running in the environment remain available:

  • Virtual machine migration
  • Multiple I/O adapter paths
  • Virtual machine load balancing
  • Fault tolerance
  • Disaster recovery tools

 

There are myriads of other great features which require greater attention and worth mentioning. I will try to address all those in my future blogs.

In today’s world many technocrats, DBAs, CIOs and CTOs are very concerned about the expenses bill on hardware procurements and also on the optimal , consolidated and cost-effective usage of the same.  I will try to get inside the problem with the help virtualization as a solution and try to build a series of blogs on virtualization (with special reference to vBlock and Oracle) . In this particular blog I will try to address the basic concepts of virtualization and its comparison with the traditional physical server architecture.

Virtualization Traditionally, operating systems and software run on a physical server. Many challenges exist in running a large number of physical servers in a typical datacenter. To run this type of infrastructure may not be efficient and cost effective in the long run. To plan and spend for the maintenance cost of this type of very large infrastructure (square footage, rack space, power, cooling, cabling, and server provisioning) are some of the problems that IT staff  and their managements are addressing on a daily basis.

Typically, there exists a one to one correspondence between a physical computer and the software that it runs. This relationship leaves most of the resources of the computers hugely idle and underutilized, leaving between only 5–15 percent (approx.) of physical server capacity in use. The cost of the space and power required to house, run and keep these systems cool can be expensive.

It is very tedious to provision physical servers as it is a time consuming process. In nonvirtualized environments time is required to procure new hardware, and get it installed in the datacenter, install and patch an operating system and finally install and configure the required applications on the same and that can take a huge time duration to get going. This process also includes many other tasks to integrate the system into the infrastructure. For example, configuring databases, servers, firewall rules, enabling switch ports and provisioning storage. Below figure shows us a basic landscape of virtualization.

cc1.png

 

Benefits The benefits of using the virtual Box vis-a-vis Physical box can be summed up in the following chart.

 

 

cc2.png

In physical environments, the OS is installed on the top of the physical hardware and while upgrading ,we need to ensure that the device drivers are in right version and installed properly as per the requirements and latest compatibility matrix. If there are incompatibilities, there may be some adverse implications in terms of business cost ,time and personnel.

Virtualizing these systems save on this cost because virtual machines are 100 percent software. The virtual machine is a set of files. A virtual machine uses standardized virtual device drivers. The hardware can be upgraded without change to the virtual machine.

Architecture The difference between the Physical and the virtual architecture can be depicted in the below diagram:-

cc3.png

The graphics shown above illustrate the differences between a virtualized and a nonvirtualized host. In traditional architectures, the operating system interacts directly with the installed hardware. It schedules processes to run, allocates memory to applications, sends and receives data on network interfaces and reads from and writes to attached storage devices. In comparison, a virtualized host interacts with installed hardware through a thin layer of software called the virtualization layer or hypervisor. The hypervisor provides physical hardware resources dynamically to virtual machines as needed to support the operation of the virtual machines. The hypervisor allows virtual machines to operate with a degree of independence from the underlying physical hardware. For example, a virtual machine can be moved from one physical host to another. Also, its virtual disks can be moved from one type of storage to another without affecting the functioning of the virtual machine.

Virtualization is the panacea for many problems that related to the CPU, memory, and networking and resources bottlenecks. Virtualization is a technology that decouples physical hardware from a computer operating system and allows us to consolidate and run multiple workloads as virtual machines on a single computer. In short ,a virtual machine is a computer that is created by software that enables us to use all the computer resources in a shared manner and enables us to run  like a physical computer which runs an operating system and applications. Each virtual machine contains its own virtual hardware, including a virtual CPU, memory, hard disk, and network interface card, which look like physical hardware to the operating systems and applications.

 

Scope for Virtualization

 

There are multiple benefits/scope for virtualization as shown below :-

 

  • CPU Virtualization
  • Memory Virtualization
  • Networking Virtualization
  • File System Virtualization
  • Server resources sharing or Virtualization

 

In the next blog I will discuss the architecture of vSphere and its allied features in greater detail. In the future blogs, I will try to take take a deep dive into the various aspects of virtualization with special reference to vSphere and oracle. Gradually I will highlight why Oracle should be used in  virtualized environment like vSphere which leverages the performance of Oracle Database and also reduces the TCO of the DB in terms of licensing costs.

 

Further Suggested Readings : Best Practices for Virtualizing Your Oracle Database – Datastores

This is a follow-up to my recent blog post EMC World 2014 - Moments (part 1), showing photos of my journey from Frankfurt in Germany to EMC World 2014 in Las Vegas, discovering the conference theme of REDEFINE, then finding my way to the EMC Oracle Booth. Part 1 rounded off with photos of the extraordinary start to the first keynote with Jonathan Martin kicking off the event playing guitar, it was then followed by Joe Tucci and David Goulden introducing EMC World 2014 and taking us through many announcements, including the DSSD acquisition, the XtremIO in-line data services $1M challenge and announcing EMC Elastic Cloud Storage (ECS, formerly Project Nile).

 

With this second blog post, my photos first highlight the exceptional presentation by Sam Marraccini “Life In The Flash Lane”! Following the theme of the Four Storage Architectures, Sam does an incredible job explaining how EMC’s portfolio of Flash-based technology addresses the needs of customers at every level of the hardware stack and in all the different storage architectures. I encourage everyone to get to see Sam’s presentation and I will update this blog entry as soon as I can link to an online video of it.

 

The remaining photos in this blog show moments where customers and EMCers alike enjoyed the event and then as all good things must, the conference comes to an end and I make my way home.

 

There are many great summaries and technology updates around the activities and announcements at EMC World 2014, the keynote sessions and links to the product announcements are already available online and many of the other technical sessions will be made available there on 28th May 2014.


Part 3 of this blog series will link to and highlight a number of the other sessions I attended and would encourage Oracle folks to view in order to appreciate the many great presentations that were available to Oracle Specialists at EMC World, perhaps next time we can organise a special track for Oracle Attendees to follow, understanding that the EMC World conference is not only for Storage professionals.

 

Enjoy the photos and online material, hope to see you at EMC World 2015 - currently planned for 5-8 May 2015 at the very same place as this year, The Venetian Resort and Hotel, Las Vegas, United States.

 

 

 

IMG_0786.jpg

The incomparable ...

IMG_0788.jpg

@SamMarraccini

IMG_0790.jpg

... Explains!

IMG_0794.jpg

Every Architecture ...

IMG_0795.jpg

... Can Leverage Flash!

IMG_0802.jpg

My personal favorite - EMC XtremCache!

IMG_0812.jpg

Attendees relaxing, ...

IMG_0814.jpg

Having fun, and ...

IMG_0817.jpg

Having a lot of fun!

IMG_0820.jpg

Who's the Waldo?!

IMG_0823.jpg

Best EMC Oracle T-Shirt!

IMG_0827.jpg

Waldo explains Oracle Re-Defined!

IMG_0867.JPG.jpgAll packed up, but not forgotten!

IMG_0886.jpg

Airports can be lonely places!

IMG_0889.JPG.jpg

Leaving Las Vegas!

IMG_0590.jpg

My Bird of Prey Awaits!

IMG_0595.jpg

12 Hour Flight from Fraknfurt!

IMG_0598.jpg

Approaching US Territory!

IMG_0600.jpg

At the Taxi Stand!

IMG_0609.jpg

The Venetian Hotel!

IMG_0618.jpg

RE-DEFI-NE!

IMG_0690.jpg

IMG_0692.jpg

Mary Bracoloni - Segway Drifting!

IMG_0695.jpg

ORACLE - Redefined!

IMG_0708.jpg

Jonathan Martin plays Guitar!

IMG_0715.jpg

Joe Tucci says Thank You!

IMG_0727.jpg

David Goulden - Re-Defined!

Slide07.jpg

Welcome to the EMC Family!

BoQKsf4IgAA6J3g.jpg

The XtremIO $1M Challenge!

Slide10.jpg

The future of Hyper-Convergence!

Bitly URL: http://bit.ly/1nhx1pY

 

Tweet this document:

Virtual Storage Zone (@cincystorage): #Oracle performance on #EMC #XtremIO. http://bit.ly/1nhx1pY

 

Follow us on Twitter:

EMCOracle.jpeg.jpg

Mark May (Virtual Storage Zone) recently posted a very nice piece on XtremIO performance with Oracle. Here is an excerpt:

 

XtremIO is EMC’s all-flash scale out storage array designed to delivery the full performance of flash. The array is designed for 4k random I/O, low latency, inline data reduction, and even distribution of data blocks.  This even distribution of data blocks leads to maximum performance and minimal flash wear.  You can find all sorts of information on the architecture of the array, but I haven’t seen much talking about archive maximum performance from an Oracle database on XtremIO.

 

The nature of XtremIO ensures that’s any Oracle workload (OLTP, DSS, or Hybrid) will have high performance and low latency, however we can maximize performance with some configuration options.  Most of what I’ll be talking about is around RAC and ASM on Redhat Linux 6.x in a Fiber Channel Storage Area Network.

Bitly URL: bit.ly/1j7zY40

 

Tweet this document:

#EMC IT: Best practices for virtualizing your #Oracle database – #VMware datastores. bit.ly/1j7zY40

 

Follow us on Twitter:

EMCOracle.jpeg.jpg

 

Related posts:

 

Darryl B. Smith with EMC IT recently published a great new blog post called Best Practices for Virtualizing Your Oracle Database – Datastores, which is part 4 of a series. Here is an excerpt:

 

First off, my apologies for delaying the last part of this four part blog for so long.  I have been building a fully automated application platform as a service product for EMC IT to allow us to deploy entire infrastructure stacks in minutes – all fully wired, protected and monitored, but that topic is for another blog.

 

In my last post, Best Practices For Virtualizing Your Oracle Database With VMware, the best practices were all about the virtual machine itself. This post will focus on VMware’s virtual storage layer, called a datastore.  A datastore is storage mapped to the physical ESX servers that a VM’s luns, or disks, are provisioned onto. This is a critical component of any virtual database deployment as it is where the database files reside.  It is also a silent killer of performance because there are no metrics that will tell you that you have a problem, just unexplained high IO latencies.

Anyway, I love Darryl's stuff and highly recommend this blog post. Enjoy.

cost-savings.jpgLast week (during EMC world) a discussion came up on Twitter around Oracle licensing and whether Oracle would support CPU affinity as a way to license subsets of a physical server these days.

Unfortunately, the answer is NO (that is, if you run any other hypervisor than Oracle’s own Oracle VM). Enough has been said on this being anti-competitive and obviously another way for Oracle to lock in customers to their own stack. But keeping my promise, here’s the blogpost ;-)

A good writeup on that can be found here: Oracle’s reaction on the licensing discussion
And see Oracle’s own statement on this: Oracle Partitioning Policy

So let’s accept the situation and see if we can find smarter ways to run Oracle on a smaller license footprint – without having to use an inferior hypervisor from a vendor who isn’t likely to help you use it to reduce license cost savings…

 

Read the entire article here.

vmworld2014.jpg

Bitly URL: http://bit.ly/vmworldvoting

 

Tweet this document:

#VMworld 2014 public voting is now open. http://bit.ly/vmworldvoting Please vote for your favorite sessions. #EMC #EMCElect

 

Follow us on Twitter:

EMCOracle.jpeg.jpg

VMworld 2014 looms, and the public voting for technical sessions is now open.

 

The following list includes EMC-sponsored sessions which are of interest to an Oracle audience:

 

 

There are a number of other EMC-sponsored sessions relating to Microsoft SQL Server and SAP HANA which are interesting, at least to me:

 

  • Session 1701 A Flash-Optimized Reference Architecture for Consolidating and Building a High Performance Virtualized Microsoft SQL Server Infrastructure on VMware
  • Session 2309 Reduce Your Business Risks and Lower Deployment Costs by Virtualizing SAP HANA
  • Session 1328 Choosing the Appropriate Storage Solutions for Microsoft Tier-1 Applications

 

Those interested in the Hadoop / OpenStack stuff might also enjoy these:

 

  • Session 1397 Benefits of Virtualizing Hadoop
  • Session 1314 Scaling Your Storage Architecture for Big Data - How the Isilon Server Fits Well with Virtual Hadoop Workloads

 

Public voting closes this Monday, i.e. May 12, at 5 P.M. PDT. Please vote early and vote often. Thanks!

Data Domain system : EMC Data Domain storage systems are traditionally used for disk backup, archiving, and disaster recovery. An EMC Data Domain system can also be used for online storage with additional features and benefits. A Data Domain system can connect to your network via Ethernet or Fibre Channel connections.

Data Domain systems use low-cost Serial Advanced Technology Attachment (SATA) disk drives and implement a redundant array of independent disks (RAID) 6 in the software. RAID 6 is block-level striping with double distributed parity. Most Data Domain systems have a controller and multiple storage units.

A Data Domain system is:

 

  • A storage system used for backup and archiving workloads that:

o   Performs high-speed deduplication to maximize storage efficiency

o   Ensures recoverability of data through integrated data integrity intelligence

o   Can replicate data automatically for disaster recovery

o   Easily integrates via Ethernet and Fiber Channel into existing backup infrastructures

  • Safe and reliable

o   Provides Continuous recovery verification, fault detection, and healing for end-to-end data integrity

 

DD Boost : EMC Data Domain Boost extends the optimization capabilities of Data Domain systems for other EMC environments, such as Avamar and NetWorker,  Greenplum, Quest vRanger, Oracle RMAN etc. DD Boost is a private protocol that is more efficient for backup than CIFS/NFS. DD Boost shares the work of deduplication by distributing some of the processing with the application host. This feature is called distributed segment processing (DSP). The DD Boost protocol enables backup servers to communicate with storage systems without the need for Data Domain systems to emulate tape. The application host is aware of, and manages replication of backups created with DD Boost. This is called Managed File Replication.

 

There are three basic features to DD Boost:

 

  1. A private protocol that is more efficient than CIFS or NFS. DD Boost has a private, efficient data transfer protocol with options to increase efficiencies.
  2. Distributed segment processing (DSP). An optional feature to DD Boost shares portions of the deduplication process with the application host, improving data throughput. DSP distributes parts of the deduplication process to the NetWorker storage node using the embedded DD Boost Library (or, for other backup applications, using the DD BOOST plug-in), moving some of the processing normally handled by the Data Domain system to the application host. The application host performs a comparison of the data to be backed up with the library and looks for any unique segments. Thus it sends only unique segments to the Data Domain system.
  3. DD Boost provides systems with centralized replication awareness and management. Using this feature, known as Managed File Replication, backups written to one Data Domain system can be replicated to a second Data Domain system under the management of the application host. The application host catalogs and tracks the replica, making it immediately accessible for recovery operations. Administrators can use their backup application to recover duplicate copies directly from a replica Data Domain system.

bb2.png

Data Domain and Oracle Oracle RMAN is a built-in tool that allows the database administrator (DBA) to easily back up and recover data in an Oracle database. RMAN handles the coordination required to ensure that transaction integrity is preserved, and sufficient information is maintained to recover the database to any appropriate point. RMAN can create backup sets that comprise as much or as little recovery information as the DBA requires but usually include information from the database datafiles, control files, and redo and archived log files. RMAN supports performing backups to a local tape drive 1 a local disk, or a NAS device, as well as integration with traditional enterprise backup applications, as shown in the below figure

 

bb3.png

Benefits of using a Data Domain system as a target for Oracle RMAN By eliminating redundant data segments inline, Data Domain systems allow many more backups to be retained than would be possible using traditional storage. In particular, Data Domain systems use a variable-length segmentation process that is extremely efficient at finding identical segments within backups of monolithic files, such as Oracle datafiles.

 

The ability of the Data Domain system to store several weeks or months of full Oracle backups enables the DBA to implement a backup and recovery scheme with great flexibility and protection while consuming a minimal amount of physical storage. The integration of the Data Domain system into an Oracle/RMAN environment is seamless

Since the Data Domain system presents itself either as an NFS or CIFS shared storage server. Oracle/RMAN already supports and documents this type of installation for effective RMAN storage.

 

If an enterprise backup software solution such as Oracle Secure Backup or EMC NetWorker is already in use, the Data Domain system can be seamlessly integrated into this environment as well. In this case, the Data Domain system can appear as an SBT_Tape device or a disk device to the enterprise backup software solution.

 

For critical Oracle environments, it is a best practice to replicate the production Oracle data to a secondary recovery location. The DBA has many options to choose from, including technologies from Oracle such as Oracle Data Guard, solutions offered by primary storage providers, and third-party solutions. Data Domain Replicator software offers extremely bandwidth-efficient replication that is also easy to deploy, enabling DBAs to leverage RMAN to provide disaster recovery capabilities for Oracle databases.

 

The primary benefit of Data Domain Replicator is the fact that only deduplicated and compressed data is transferred across the network. Because deduplication is happening inline, replication takes place while the RMAN backup process is still active. As the RMAN backup process proceeds, the unique segments and metadata representing each file in the backup set are queued for replication to the remote site, -to- many cases, replication is completed within a short period of time after the initial backup completes.

 

References For additional information, see the following:

 

EMC Data Domain Family products and deduplication technology http://www.emc.com/products/family/data-domain-family.htm

EMC Data Domain solutions for Oracle http://www.datadomain.com/solutions/oracle.html

EMC Backup, Recovery, Archive solutions for Oracle http://www.emc.com/solutions/samples/oracle/backup-recovery-archive-oracle.htm EMC

Solutions for Oracle http://www.emc.com/solutions/application-environment/oracle/index.htm

EMC Data Domain Global Deduplication Array http://www.emc.com/products/detail/hardware/data-domain-global-deduplication-array.htm

EMC Data Domain Boost software http://www.emc.com/products/detail/software/data-domain-boost.htm

White Paper H11798 - Oracle RMAN Best Practices with Data Domain

This is a translated blog from the original blog of my colleague Mr. Makoto Miura (Makoto-san). This blog was also reviewed by Makoto-san and so I convey my utmost gratitude to him.

 

From my previous blog, we have understood the features and the benefits of storage based backup. Will not we now learn about some important points while using ASM? This time I would like to write an ASM related blog which is in continuation to my previous post Foundation of the storage based backup related to the Database.

 

Though ASM can be used both with block and NFS storage, I would like to talk about the block storage first. Of all the customers that I have worked with in Japan, I found that the majority of them are using block storage. I know of only 1 customer who uses NFS. In your environment, which storage option is most frequently used?

 

  • At the time of mounting the copied ASM DG to the production server, kindly rename the ASM DG.

 

In earlier versions of Oracle, we could start the multiple ASM instances on a single DB server box individually.  At the time of mounting the server for the purpose of copying ASM DG, it was possible to use the same name of the ASM DG. But in the new version of Oracle viz. Oracle 11gR2, in a single DB server we can start only ONE ASM instance. For this reason, before mounting the copied ASM DG, the name of the ASM DG needs to be changed. If we use EMC Replication Manager then the process of renaming the ASM DG name gets automated.

 

  • Maintaining the Consistency of the ASM DG

 

It is very important to maintain the consistency within DG when we build the DG from multiple LUNs. The metadata information (related to the ASM DG) is maintained within the DG itself. So, at the time of rebalancing, when hot backup mode is used, it is very important to maintain the write-order consistency in the LUN unit of the ASM DG. Within the LUNs of the EMC Storage, multiple logical groupings can be created, thereby the consistency is also maintained. This happens not only in high-end storage, but also in the mid-range storage. Yes, we can achieve this.

 

  • Important Points during the usage of ASM’s Failure Groups

 

If some users want to perform Data mirroring between different storage boxes while using ASM using the Failure Group, the data gets scattered between storages. In the below figure, the dispersion of data is getting depicted for ASM-DG#1 to two failure groups.

 

bb1.png

 

The above figure shows the self-setup function of ASM’s failure group creation.

 

Now, which process will be appropriate when we want to setup ASM-DG#1by storage copy?

 

To be frank, in the above setup oracle supports in maintaining the integrity of the whole setup and performs the copying operation. In short, one failure group’s copy looks insufficient here.

 

Let’s look into the various failure scenarios:-

 

  1. Within the failure group , one failure group got crashed.                                                                                    At this situation, the crashed Failure group is restored from the surviving failure group/s. Here storage based copy cannot be used.
  2. Within the failure group , two failure groups  got crashed.

 

That type of situation does not occur normally but even if we restore 1 failure group by the storage copy mechanism,  we are not sure whether Oracle DB can  be restarted or not as Oracle (formally) has not provided any guarantee for this operation. In the above picture, Oracle supports the storage copy only when  we create both Replica#1 and Replica#2 with the same timing and the consistency should also be maintained.  To recap, when we use storage based copy with ASM failure groups, we must take copy of all failure groups with consistency.

 

  • Why the customer wants to do ASM mirroring by using failure groups?

 

I feel the true purpose is to improve the reliability of the DB system by having redundant storage infrastructure.  In this situation, we can propose many alternatives, such as storage based mirroring between storage boxes; Appliance based mirroring (RecoverPoint) and Storage virtualization (VPLEX).

pollack.jpg

Bitly URL: http://bit.ly/convergedinfrastructure

 

Tweet this document:

Soooo, what's a converged infrastructure, exactly? http://bit.ly/convergedinfrastructure #EMC #EMCElect

 

Follow us on Twitter:

EMCOracle.jpeg.jpg

I have been looking at a lot of stuff relating to the notion of a converged infrastructure. The essential idea is this: The legacy platform, i.e. a single-image, physically booted server, typically attached to a SAN, is hopelessly outdated. This design is essentially 30 years old, and does not take into account many interesting recent technology changes, notably:

 

  • Virtualization
  • Deduplication
  • Flash storage, including both PCIe-based flash and SSDs

 

Thus, the converged infrastructure platform looks dramatically different from the legacy platform. Essentially, a converged infrastructure consists of a large cluster which combines storage and compute into a single layer. A clustered I/O subsystem is extended across all of the nodes in this cluster, and all I/O to this shared storage pool is reflected on all nodes. This is either implemented in block or file (NFS), depending on the vendor involved. All of the storage hardware is direct attached. No SAN, in other words. Thus, a converged infrastructure is designed for data center consolidation, similar to the virtualization hypervisor market. Some of the converged infrastructure vendors run natively on a hypervisor, and some do not. More on this later.

 

The converged infrastructure market is crowded, that's for sure. It would be difficult in a blog like this to cover all of them, so I will focus on three:

 

 

And, of course, my focus is on the Oracle space. As usual, I think in terms of running Oracle as efficiently and inexpensively as possible. Also, in terms of hypervisors, I will only address VMware. (I have no technical exposure to Hyper-V or KVM.)

 

Starting with Nutanix, the architecture is best described in the Nutanix Bible by Steven Poitras. I have slightly reworked one of his graphics to reflect my bias, again Oracle on VMware:

 

nutanix architecture.png

 

Essentially, all of Nutanix's IP runs in a VM, referred to as the CVM. The hypervisor and the CVM boot off of a small partition. Then the CVM connects to the other nodes in the Nutanix cluster, and assembles a clustered file system, which is published out as an NFS export. The hypervisor then mounts this export as an NFS datastore. From there, all user space VMs (including Oracle VMs) are booted off of .vmdk files on the NFS datastore. All block-level I/O (other than boot) is handled by the CVM, which by-passes the virtualized controller and uses the native LSI driver with direct path I/O. (thanks to Michael Webster for his correction on this point).

 

Again, this is from the VMware perspective. The architectures for other hypervisors are different. But I digress.

 

Moving to SimpliVity, my source for their technical stuff would be the OmniCube Technical Deep Dive. (At 10 pages, this deep dive is not quite so deep as I would prefer, for sure.) The SimpliVity architecture is very similar to Nutanix, in most respects except one: Simplivity adds a piece of hardware which they call the OmniCube Accelerator Card (OAC). Otherwise, the architecture diagram for Simplivity looks just like Nutanix:

 

simplivity architecture.png

 

Again, all of SimpliVity's IP runs in the OVC VM, other than the OAC itself, of course. The OAC is a custom-built PCIe card, which among other things acts as the I/O controller. Like Nutanix, the OVC exports an NFS mount, which ESXi mounts as an NFS datastore. From there, all user space I/O, including Oracle, runs through the .vmdk layer within the hypervisor.

 

Now, looking at ScaleiO, the architecture is dramatically different from either Nutanix or SimpliVity. First of all, Nutanix and SimpliVity are both custom-built hardware platforms. ScaleiO is a piece of software. It is designed to be layered on top of a normal, legacy platform, and provide a slick, easy path to a converged infrastructure. Specifically, ScaleiO does not require the use of a hypervisor, and thus can run in a physically-booted context. This is one of ScaleiO's main advantages over hypervisor-based converged platforms like Nutanix and SimpliVity.

 

ScaleiO consists of two major components: The ScaleiO Data Client (SDC) and the ScaleiO Data Server (SDS). In a Linux context (again, the only OS I care about deeply), both the SDS and the SDC are implemented as kernel loadable modules, similar to device drivers. The SDS manages the local storage hardware, connects with other nodes to the ScaleiO cluster, performs cache coherency, etc. The SDC then connects to the SDS, which appears to it as a SCSI target. Thus, the SDS publishes storage objects to the SDC, which the local OS sees as normal SCSI LUNs. From there, the local OS simply performs normal I/O.

 

The stack diagram for a physically-booted ScaleiO cluster node could not be simpler:

 

scaleio phys architecture.png

 

ScaleiO can also be run in a virtualized context. In this case, predictably, ScaleiO looks very similar to Nutanix or SimpliVity, in that it has a controller VM as well, called the ScaleiO VM (SVM). This SVM runs both the SDS and the SDC. All I/O is channeled through the SVM. However, everything in ScaleiO is implemented in a block, rather than file, manner. Thus, the ESXi hypervisor sees an iSCSI target which it converts into a VMFS file system, rather than using NFS. (The SVM provides the iSCSI target for this purpose.)

 

Here is how ScaleiO looks in a virtualized configuration:

 

scaleio virt architecture.png

 

The other interesting thing about ScaleiO is that it allows you to run either in a converged or a diverged manner. Since the client (SDC) and server (SDS) are separate components, you can run them on separate hardware, effectively turning the ScaleiO server cluster into a storage array. See the following graphic for an example (thanks to Kevin Closson for this):

 

scaleiodivergedmodel.png

 

Of course, you can also run ScaleiO in a converged manner, in which case the platform looks very much like Nutanix or SimpliVity (with the exceptions noted).

 

Now, looking at each of these architectures in the context of running Oracle, it appears that ScaleiO has the obvious edge. This is because:

 

  • Both Nutanix and SimpliVity require you to virtualize in order to run on their platform. ScaleiO does not. Even the most ardent proponent of virtualizing Oracle (and I would certainly qualify on that score) would want to maintain the option to run on bare metal if necessary. Adopting a platform which requires the customer to virtualize all Oracle workloads is not going to work for many Oracle customers.
  • The use of a VMware NFS datastore as the primary container for Oracle storage is problematic, especially in an Oracle context. While I was with the EMC Global Solutions Organization, we tested VMware ESXi-mounted NFS datastores for Oracle datafiles. This configuration had a huge performance impact, relative to either normal NFS (i.e. directly mounted on the guest OS, and using Oracle Direct NFS Client) or block-based VMFS, using conventional SAN storage. There is no reason that this would be any different in a converged context. Think about it. The code path for an I/O on an Oracle ASM diskgroup which is being stored on a .vmdk file which is, in turn, on an NFS datastore is enormously long. Compare that to ScaleiO, especially in a physically-booted context, where the I/O code path is no more lengthy than using a normal device driver!
  • The diverged approach for ScaleiO is arguably custom-built for Oracle. I have made a pretty good living for the last (approximately) 17 years of my life by understanding one thing: Oracle is expensive, and therefore the Oracle-licensed CPU is the single most expensive piece of hardware in the entire configuration. Offloading any workload from the Oracle-licensed CPU onto a non-Oracle licensed CPU is typically a very profitable decision. (Arguably, this is why Oracle database servers adopted storage array technology so widely: By offloading utility operations like snaps, test / dev cloning, data warehouse staging, backup, and the like onto a storage array, the customer preserves the precious Oracle-licensed CPU to do what it does best: Run queries.) Both Nutanix and SimpliVity require the customer to run in a converged manner, thus using the Oracle-licensed CPU (on the ESXi host in this case) to run storage I/O operations. That's wasteful of that precious Oracle-licensed CPU. Thus, it is entirely possible that a fully converged infrastructure may be a poor fit for Oracle, because of simple economics. By enabling a diverged configuration (i.e. looking more like a traditional SAN storage array) ScaleiO neatly optimizes the Oracle CPU, while preserving many of the advantages of a converged platform.

 

Of course, it remains to be seen how this all pans out. It's early. For now, though ScaleiO is looking good to me.

Filter Blog

By date:
By tag: