March 2017: Introducing IBM Spectrum Control v5.2.13

IBM Spectrum Control offers a single pane-of-glass to monitor and manage multi-vendor networked storage.  The latest release Spectrum Control v5.2.13 now offers the following enhancements:

Enhanced support for Dell/EMC storage monitoring

  • In addition to Spectrum Control, IBM Storage Insights, our SaaS implementation also monitors Dell EMC storage specifically VMAX, VNX and VNXe. We’ve added support for both block and file for Dell EMC VNX and VNXe as shown below:

sc-for-emc-vnx-and-vnxe

  • We now provide visibility into volumes on Dell EMC storage (VMAX, VNX, VNXe) that are thin provisioned, compressed or even when they are meta volumes or private volumes as shown below.

sc-and-emc-meta-volumes

  • If you had created alert definitions in previous versions of Spectrum Control they can be migrated over to Spectrum Control 5.2.13 as shown below.

SC and alert definitions.jpg

More insights from storage that is external to Spectrum Scale

  • If you are a user of IBM Spectrum Scale and decide to move cold data from the Scale file system into external storage (say IBM Cloud Object Storage) wouldn’t it be nice to know that you have enough local storage capacity in the Scale file system before you recall archived data from IBM COS? This is now possible by merely looking at the shortfall column – if the value in this column is 71.69% you have a red flag informing you that you don’t have enough local storage space so you should refrain from issuing a recall.

sc-and-shortfall-view

Enhancements for IBM object storage (COS)

  • While IBM COS has its own device manager dsNet to manage multiple storage devices from multiple vendors it is convenient to have a single pane of glass like that provided by Spectrum Control which can drill down into HDD, SSD and flash usage in IBM COS Slicestor® nodes as shown below:

sc-and-ibm-cos-nodes

Enhancements to replication capacity planning

  • We now provide a way to identify your risk exposure by using Spectrum Control to identify primary data that hasn’t been protected by replication.
  • You can report on storage capacity consumed by copy data & replication. The following visual shows you how we would display information about vdisk mirror for Spectrum Virtualize or Storwize.

sc-and-copy-data

  • Within the Copy Data view you can also view HyperSwap relationships as shown below.

sc-and-hyperswap

For a detailed list of what’s new in Spectrum Control 5.2.13 please refer to this link.

For  the latest storage inter-operability matrix for Spectrum Control 5.2.13 please refer to this link.

Dec 2016: Introducing IBM Spectrum Control v5.2.12

IBM Spectrum Control offers a single pane-of-glass to monitor and manage multi-vendor networked storage.  The latest release Spectrum Control v5.2.12 now offers the following enhancements:

Capacity and performance monitoring of the IBM FlashSystem A9000 & A9000R

If you are a Managed Service Provider (MSP) offering cloud based services based on IBM storage like XIV or FlashSystems, you may wish to consider the FlashSystem A9000 which offers an all-flash enclosure node with three grid controllers.  If you need to scale even further the FlashSystem A9000R which is the rack version, supports multiple flash enclosures and multiple grid controllers.

capacity-management-of-a9000

While the FlashSystem A9000 has its own user interface, Spectrum Control provides significant value when you need a central pane of glass to manage the A9000 alongside other IBM storage (like Spectrum Scale, Spectrum Virtualize, Storwize, DS8K, XIV) and non-IBM storage (EMC, NetApp and others).  Spectrum Control v5.2.12 supports:

  • Capacity management: Monitors capacity in GiB at the pool level and also displays the savings you achieve from data reduction techniques (block level data de-duplication and in-line compression).
  • Performance monitoring: For Fibre Channel ports and storage volumes

Storage consumption reports by different consumers

As an enterprise IT dept offering internal cloud services you may have a need to create reports on block storage use by different consumers (who may be the application, department, hypervisor or physical server).  You may have a need to identify storage consumption by storage tier or by copy data.  Spectrum Control v5.2.12 supports this and gives you the option to use default costs or to assign custom costs to block storage, file storage and storage capacity used for copies.

storage-consumption-by-consumer

Copy data reporting

As a storage manager you may want to know how much of your data in networked storage is primary data versus mirrored data.  Spectrum Control v5.2.12 reports on copy data regardless of the actual copy mechanism (Metro Mirror, Global Mirror, Global Mirror with Change Volumes, FlashCopy, Global Copy).

If you encounter performance problems and you suspect that the way replication has been setup is a probable cause, then you can use Spectrum Control v5.2.12 to drill down and get a list of consistency groups,  or even a list of all relationships that are part of a consistency group.

copy data reporting by consistency groups.png

copy-data-reporting-by-flashcopy

Reporting on storage capacity in pools “external” to Spectrum Scale

What if you need to understand the storage capacity in internal as well as external pools of Spectrum Scale file systems?  Spectrum Control v5.2.12 now enables this.

 reporting-on-external-pools

A single pane of glass to monitor IBM Cloud Object Storage along with other storage

As an enterprise or MSP already using a variety of IBM storage you may have invested in IBM Cloud Object Storage (COS) to get an archive tier for less frequently used “cold” data.  Spectrum Control v5.2.12 complements dsNet Manager and helps you improve your risk posture by identifying which IBM COS site cannot tolerate an IBM COS Accesser node failure.    It also gives you a way to identify quota (hard/soft) compliance at the IBM COS vault level.

Greater choice on when tiers are analyzed and pools balanced

As an enterprise customer whose workloads spike during normal business hours Monday-Friday and then taper off on weekends you may want to isolate storage tiering analysis only to business hours.  This is now possible with Spectrum Control v5.2.12

choice-on-scheduling-tier-analysis

For a more detailed list of what’s new in Spectrum Control v5.2.12 please refer to this link.  If you have thoughts or suggestions on how Spectrum Control could add more value to your storage management tasks please feel free to contact me.

August 2016: Introducing IBM Spectrum Control v5.2.11

IBM Spectrum Control offers a single pane-of-glass to monitor and manage multi-vendor storage.  V5.2.11 now introduces the following enhancements:

Performance monitoring of IBM FlashSystem 840 and FlashSystem 900

Are you perhaps a Managed Service Provider (MSP) or an enterprise customer with a need to accelerate performance for applications like Oracle, DB2, SQL Server, MySQL, Sybase, NoSQL?  If you are running applications requiring response times in the 100 microsecond range (as opposed to milliseconds) and you have storage services provided by the application, then FlashSystem 900 is your logical choice.   This platform features 90us/155us read/write latency and has the ability to sustain 1.1 million random read 4 K IOPS.

IBM Spectrum Control v5.2.11 adds performance monitoring for the FlashSystem 840 and FlashSystem 900  – specifically metrics like response times, I/O rates and data rates.  This can be done at the level of Fibre Channel ports or at the drive level.   It doesn’t report on response time of individual volumes.

perf chart for FlashSystem 840

IBM Cloud Object Storage as an external pool for IBM Spectrum Scale

What if you have an HPC grid built on IBM Spectrum Scale (previously GPFS) and need to know which file systems in your IBM Spectrum Scale cluster are using cloud storage?  While the element manager in Spectrum Scale can give you visibility into file data residing in internal pools,  Spectrum Control goes beyond this to give you visibility into files residing in external pools (on IBM Cloud Object Storage, Spectrum Protect or Spectrum Archive).  You can drill down to view storage capacity at each site of IBM COS on a per-vault basis.  You can identify how much failure can be tolerated within IBM COS in terms of storage node or access node going offline.

Complex alerts and custom multi-conditional alerts

While alerts are always useful, sometimes there can be too much of a good thing.  Consider this scenario: You have legacy EMC, HDS and other storage being virtualized by IBM SAN volume controller.  You have 30 host applications that receive volumes from the IBM SAN volume controller.  10 of these are critical apps which require that latency be <= 10 msec.  Why have a deluge of alerts whenever latency crosses 10 msec on all volumes especially if 20 of the apps using these volumes can tolerate higher latencies up to 20 msec?  To address this scenario, v5.2.11 gives you the ability to set alerts to be informed when latency >= 10 msec for only those volumes that are used by the latency sensitive apps.  This makes the alerts more meaningful and actionable.

The following is an example of how you could set an alert to inform you when all volumes associated with tier 1 storage have an overall response time that is >= 6 msec.  This helps you cut out false-positives which would have been generated if the alerting occurred on all tiers esp if tier 2 and tier 3 storage are used by less critical (less latency sensitive) host apps.

custom alert

For performance monitoring if you look at I/O rate alone you may not be accounting for large vs small packets.  However if you combine multiple conditions (I/O rate, response times, data rates) you can draw more meaningful conclusions.  With this in mind, Spectrum Control v5.2.11 now supports multi-conditional alerting.

Chargeback reports from the web UI without requiring Cognos

Whether you are an MSP or an enterprise, the ability to produce chargeback reports is critical when you have external or internal consumers of shared storage. As the storage administrator you get to set the cost per tier on a cost per TB basis.  The following is typical of a storage consumption chargeback report that could be generated to show shared storage (block or file) consumed by a department, group, application, physical server or hypervisor:

charge_back_reports_cfg

Automatic probe scheduling

As a storage admin you may not want to individually manage when Spectrum Control connects to your storage subsystems especially if you have 100s of storage systems and 1000s of servers.  To address this, v5.2.11 introduces automatic probe scheduling.  Now  all the storage administrator has to do is specify when probes should not be sent (perhaps during normal business hours of 9 am to 5 pm) and Spectrum Control will use analytics to load balance the probes over an approved time window.  This feature is applicable for storage systems as well as for hypervisors.

License compliance views for audit purposes

In the past, you may have had to manually determine your licensed storage capacity for each storage system (by capacity or by enclosure) for audit purposes.  Now v5.2.11 offers a way to view your license compliance by storage capacity in tebibytes (TiB) and storage capacity units (SCU).

Learn more by reviewing the guided demo of our SaaS offering “Storage Insights”  Consider a free 30-day trial of Storage Insights.

Consider upgrading your on-premise version of Spectrum Control to v5.2.11: Downloads available here.  

 

IBM DS8000 performance analysis using Spectrum Control

What do most of the top 20 banks in the world have in common? They all require “always on” 24×7 highly reliable networked storage for business critical applications, greater than 6-nines of reliability, replication with a Recovery Point Objective (RPO) of 3 to 5 sec.  This is why most of them have standardized on the IBM DS8000 line of tier 1 storage along with IBM z systems for their demanding workloads.  Outside the banking industry numerous organizations use the DS8000 attached to open systems hosts.

How does IBM Spectrum Control add value to customers deploying DS8000 with mainframe or open system hosts?  If you have multiple DS8000 systems across different locations,  a view like the one below gives you an idea of the location, firmware levels, volumes on each DS8000 that are in a synchronous relationship (Metro Mirror), volumes that are in asynchronous relationships (Global Mirror).  If you notice an error condition, you can double-click on that row in the panel to get more details of the error

Enterprise wide monitoring

While the element manager for the DS8000 gives you insights into that particular storage array, what if you want to compare performance across multiple DS8000?  This is where Spectrum Control shines.  The following panel shows you performance in operations per sec over a 16 hour time window.

compare performance

If you should happen to have some legacy HDS VSP or an EMC VMAX in your data center, Spectrum Control will also display similar performance views for those arrays.

What if you wish to look at performance across sub-components like nodes, pools or volumes?  The following panel shows the visibility you gain using Spectrum Control.

perf across subcomponents

Considering that Spectrum Control monitors over 100 performance metrics on the DS8000, you may want to set a threshold to generate a performance alert if the disk utilization goes above 80%.  The following panel shows you how this would look.  You can also view details of volumes which contribute to this scenario.

performance alert

What if you want to identify when volumes were created or deleted in a specific pool?  The following view from Spectrum Control shows historical capacity for pools during a 1 month period.  You will notice that for pool P5, volumes were deleted on May 22 and possibly new ones created on May 26.  We also infer that used capacity (depicted by a solid line) is still below available capacity (dotted line).  This level of insight helps you with capacity planning.

historical capacity

With the goal of maximizing your Return on Investment (ROI), you may wish to identify wasted storage capacity with the intention of reclaiming it.  The following panel shows you how you would achieve this.  If there has been no recent I/O on volumes within a pool it is reasonable to assume that this capacity may be unused hence reclaimable.

detect wasted capacity

In conclusion, if you have a need for enterprise wide performance analysis, alerting, tiering analysis or capacity monitoring then Spectrum Control is worth considering.

Consider a free 30 day no risk evaluation of Storage Insights, if you find the insights you gain valuable you can either purchase a subscription to have IBM manage your storage from the cloud or purchase the full blown Spectrum Control product for your on-premise  use.

May 2016: Introducing IBM Spectrum Control v5.2.10

IBM Spectrum Control is Storage Resource Management (SRM) software which offers a “manager of managers”, a single pane of glass to monitor and manage multi-vendor storage (block, file and object) across your data center.

Version 5.2.10 released this month, introduces the following additional benefits to storage management teams who may use it to manage multi-vendor networked storage:

  • Faster time to problem resolution
    • The ability to export performance data directly from the web browser to a compressed file which could be sent over to IBM support if you need help analyzing a performance problem.

fig1

fig2

  • Performance monitoring of outliers
    • Assume you are using IBM Spectrum Virtualize or Storwize to virtualize legacy multi-vendor storage from EMC, HP, and HDS. You may have an “acceptable range” for performance metrics and may want to be alerted if one of the metrics like “read response time for an I/O group” occasionally falls outside your acceptable range.  This is now possible with Spectrum Control as shown below:

fig3

fig4

  • Capacity planning when thin provisioning is involved
    • You may want to increase storage utilization hence you use “thin provisioning” to oversubscribe your existing networked storage.  However thin provisioning assumes that you will add storage to your thin provisioned LUN in time before the application crashes.  To do this you need a way to predict when you’ll hit the upper ceiling where thin provisioned storage must reconcile with your actual physical storage capacity.  Spectrum Control now gives you a way to do this by looking at past capacity trends and charting historical capacity of thin provisioned volume growth.
  • Historical capacity charting
    • You may want to improve overall storage system performance and hence you enable IBM EasyTier. Now you have the option to do historical capacity charting of tier distribution in EasyTier pools.
    • You may be using IBM Spectrum Scale (previously called GPFS) and may want to use Spectrum Control for capacity planning of Scale clusters. You now have the option with 5.2.10 to do historical capacity charting of Spectrum Scale file systems, pools and file sets.
  • Greater control over the operation of your Spectrum Scale clusters
    • For your Spectrum Scale cluster you as the storage administrator may want to schedule the load placed on it when Spectrum Control collects information regarding the size of Spectrum Scale (GPFS) snapshots. Now you have the option to decide when/if you want to enable the collection of snapshot size information from Spectrum Control
  • Consolidated storage reporting across multiple data centers
    • You may want a network-wide perspective of storage across data centers managed by multiple Spectrum Control instances. Now the roll-up reporting feature introduced in 5.2.10 gives you the ability to roll up capacity reports across all of these instances.

fig5

  • Streamlining of SAN monitoring
    • Unlike in the past where you had to choose between SNMP or Storage Resource Agent (SRA) or Storage Management Initiative (SMI) to monitor your Brocade SAN, now we have streamlined it down to one standard way to access your Brocade switches namely SMI. We recommend that you connect to your Brocade SAN switches via Network Advisor version 11.1 (with its own embedded SMI agent) or later.

In conclusion these are just a few of the many exciting new developments to come for Spectrum Control.  In addition, if you are interested in learning more about our SaaS offering, check out the IDC blog about Spectrum Control Storage Insights written by our guest blogger Jingwen Li from IDC.

 

Value of Spectrum Control to Spectrum Accelerate

Are you an enterprise customer or cloud provider looking for a distributed tier 1 storage offering consistent performance for use with OLTP, Oracle, SAP, Exchange or Web 2.0 workloads?  Look no further than IBM Spectrum Accelerate.  Spectrum Accelerate is based on what was previously IBM XIV and is ideally suited for customers with PB scale needs but just a minimal storage staff available to manage it.  Don’t take my word for it but hear why a customer like Silverpop (vendor of marketing automation from the cloud) chose Spectrum Accelerate over  VMware VSAN and EMC ScaleIO for a high-write Oracle environment.

Now you may wonder – what value would I get from IBM Spectrum Control considering that Spectrum Accelerate already includes Hyper-Scale Manager?  Spectrum Control is designed to monitor and manage multi-vendor storage including Spectrum Accelerate.  It provides storage teams with a single pane of glass to view and report on health, capacity, performance, and SAN connectivity of the entire storage environment, as well as alerting on issues that occur in the environment.  Here is a view of Spectrum Control monitoring Spectrum Accelerate alongside other IBM storage xiv1

Using this tool, you can identify health issues and use the capacity views to identify when your environment runs out of storage.  In addition to viewing capacity you can also chart it historically for planning purposes.

The performance view shown below helps you identify that Spectrum Accelerate is pushing the most IO in the environment: xiv2

Even though it is pushing the most IO, it has the best overall response times as shown below.  xiv3.png

As a storage consumer, perhaps you have a need to generate chargeback reports?  To help you Spectrum Control supports consumer modeling, enabling you to create models of how your servers, applications, and departments are using shared storage.

Perhaps you have a need to understand how capacity trends over time for an application (say a billing app) when it is using multiple volumes in a Spectrum Accelerate solution?  The following screen shot demonstrates such capacity trends for an application that uses 5 volumes in Spectrum Accelerate. xiv4

Perhaps you have a need to identify response times for the application when it hits those storage volumes?  In the above example all volumes have a response time of less than 6 msec per operation.

For chargeback perhaps you need to contrast the storage capacity consumed by one host application versus capacity consumed by other host applications?  The following screen shot demonstrates how this would appear.

xiv5.png

From the perspective of alerting, Spectrum Control can create over 100 different alerts for Spectrum Accelerate.  From a reporting perspective Spectrum Control provides over 70 Cognos reports and gives you the option to create your own custom reports. From a cost savings perspective, Spectrum Control helps you identify storage volumes that have been orphaned by decommissioned applications on host servers, so you may re-purpose those volumes.  This allows you to defer additional storage purchases.

Like what you hear and want to give this a test drive?  Check out our 30 day no risk evaluation of Storage Insights, if you find the insights valuable you can either purchase a subscription to have IBM manage your storage from the cloud or purchase our comprehensive on-premise offering that is Spectrum Control.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Value of Spectrum Control to Spectrum Scale

Today if you are a customer in a sector like financial, retail, digital media, biotechnology, science or government and you use applications like big data analytics, gene sequencing, digital media or scalable file serving, there is a strong possibility that you are already using IBM Spectrum Scale (built upon General Parallel File System or GPFS).

Spectrum scale

A question foremost in your mind may be: “If Spectrum Scale has its own element manager – the Scale GUI, what would I gain from using Spectrum Control?”
The Spectrum Scale GUI (available for version 4.2 and later versions of Spectrum Scale) focuses on a single Spectrum Scale cluster. In contrast, the Spectrum Control GUI offers a single pane-of-glass to manage multiple Scale clusters (including version 4.1.0 of Spectrum Scale), it gives you higher level analytics, a view of relationships between clusters, the relationships between clusters and SAN attached storage.  In future, we expect to extend this support to Spectrum Scale in hybrid cloud scenarios where Spectrum Scale may be backed by say Amazon S3 or IBM private cloud storage.

4 primary areas where Spectrum Control adds value today:

Spectrum Control for Scale

  • Performance monitoring: If you have multiple Spectrum Scale clusters you can use Spectrum Control to manage cross cluster traffic since Spectrum Control charts I/O rates and response times. You can also monitor the performance of the file systems and nodes that make up each cluster. Since Spectrum Control can monitor block storage devices, you can monitor the performance of SAN attached volumes that are backing the NSDs in the cluster, as well as the fabrics used to connect the NSD servers to the block storage.
  • Capacity management: If you have say an IBM FlashSystem providing block storage to your Spectrum Scale clusters, you can monitor these block volumes backing the Scale NSDs. You can identify when you will run out of inodes or when a file system runs out of space. You can view the size and age of snapshots in a file system perhaps to identify candidates for deletion and space reclamation. You can view file system relationships, quotas, snapshots, NSDs being used to build a pool, the number of nodes using the file system, compare pools across the Scale cluster.
  • Alerting: Spectrum Control enables meaningful alerts by nodes, NSD, file systems, file sets. You can identify (though not alert upon) which file sets will run out of available inodes, identify filesets not linked into their file system, identify (though  not alert upon) which filesets have reached soft/hard quota limits. As an example, you may want a warning alert notification sent to your operations team when the used space under Spectrum Scale crosses 80%, but when the used space crosses 90% you may want a critical alert notification sent to your storage administrator. Spectrum Control handles all of these scenarios.
  • Showback & chargeback: Spectrum Control offers the ability to categorize storage resources used by Spectrum Scale into “application groups” and “department groups” which in turn is a key enabler of showback and chargeback.  The benefit is that now a Spectrum Scale administrator can identify issues that impact specific applications.

If you like what you hear so far but don’t have an on-premise version of Spectrum Control in-house, check out our Spectrum Control Storage Insights offered as SaaS and give it a risk free 30 day test drive.