Welcome to the latest Ask the Expert session on the ViPR Community! This discussion will focus on the recent test study performed by Principled Technologies examining how EMC ViPR customers can reduce operating expenses (OX) with EMC ViPR Controller and ViPR SRM products. Please join this discussion and send us your questions until April 17th, 2015.

If you have not had a chance to read the study yet, please visit Test Report: Reducing Operating Expense with EMC ViPR Controller and ViPR SRM.

Niki

(ViPR Community Manager)

Our ATE event is now live. You may post your questions to our SMEs. Let's keep this discussion informative and respectful. Who has the first question? Cheers!

Thanks Niki. Sorry in advance for the long answer but here is the approach and scenario/test environment we used in our study. 

The goal was to analyze potential OPEX savings from deploying ViPR Controller and ViPR SRM, so Principle Technologies researched salary information from publicly available sources for a variety of IT functions including analysts, administrators, specialists, directors- and C-level roles. We then used this data as we created our example ITSM change management process, which defines how an organization would manage customer requests to ensure minimal risk to the operating environment. In addition to ticket handling and transfers, we evaluated the OPEX costs associated with the Change Advisory Board (CAB). The CAB is a service management layer that reviews each requested change for its impact and risk, and approves and schedules changes for implementation.

We conducted customer interviews that provided us with real-world experiences for our framing scenario, and supporting

data validating our calculations. In addition, we leveraged the hands-on testing from our previous EMC ViPR Controller work where applicable.

We considered how a large enterprise, with multiple data centers each containing storage platforms of various age and from disparate vendors, would operate. In our example, this large enterprise has thousands of internal and external customers who require IT provisioning, data protection, and management services.  Our organization has six different regional data centers geographically dispersed over a wide expanse, each with a variety of active block & file storage platforms. Each data center has three shifts of employees with at least one storage, network, server, and virtualization administrator on shift at all times.

To calculate OPEX savings for our enterprise, we researched the salaries of storage, network, and server administrators; help-desk staff; IT and executive management; and other supporting full time employee costs using publically available sources.

Our test environment, in the Principle Technologies lab, simulated various storage platforms a large organization might encounter. To show how EMC ViPR Controller or ViPR SRM can manage these various platforms, we used an EMC VNX 5500, an HP EVA 4000 physical array, NetApp and EMC Isilon simulators running as virtual appliances on a dedicated infrastructure virtualization host, and a non-supported Dell EqualLogic 6110 physical array.

To manage the HP EVA 4000, we set up a Windows Server 2008 R2 x64 server with HP P6000 CommandView 10.1 and physically connected the server to the SAN environment. As it was otherwise unutilized during testing, we configured the EMC SMI-S provider, which allowed us to discover and manage our EMC VNX 5500, and the HP SMI-S provider, which allowed us to access the EVA 4000 using non-HP tools such as ViPR SRM on alternate ports.

Within this environment, we executed hands-on testing that supplemented our previous findings and validated those findings with the latest versions of ViPR Controller and ViPR SRM.

Our first question is for JohnMcMains:

What customer scenario and test environment was used by Principled Technologies for this study?

Thank you John.

Can you also comment on what the provisioning experience was like with ViPR Controller and what savings the study revealed for storage provisioning?

You bet. ViPR Controller delivers uniformity to the storage provisioning process, which can help eliminate delays due to human error. Regardless of the vendor or model of storage from where your administrators provision, ViPR Controller provides a single common interface, driven by pull-down menus, for selection of the attributes your users need for their storage profile. It provisioned with a just few clicks in our case.

The organization in our example could save up to $64,320 per year in administrative costs for those tasks, which means up to a 73% reduction in storage provisioning costs with ViPR Controller.

Good stuff John!

To add a little extra color on how ViPR is able to help with those savings, below is an example of provisioning a block volume with ViPR Controller to a host.

ViPR Controller automates the following steps to provide to provisioning a block volume with ViPR Controller to a host:

1. Run through a series of preliminary checks including checking that zoning exists between the host initiators and storage adapters.  In this example zoning already exists.
2. Create the Block Volume on the physical array that is obscured to the user from the Virtual Pool on the Virtual Array per any selected requirements
3. Create access from the Physical array Block Volume to the Host
4. Connect to the host, get required locks, and rescan for disks
5. Connect to any new disk, format and mount the disk, and names it

Here’s what the provisioning order screenshot looks like:

Thank you Vinay and John.

Can you please elaborate on how SAN zoning process is handled within ViPR Controller?

Sure thing. ViPR Controller can eliminate the need for storage administrators to perform any SAN zoning tasks for Brocade and Cisco SAN infrastructure. In our testing we discovered and registered the fabric management switches, compute hosts, and storage arrays within ViPR Controller, and ViPR Controller created these logical connections automatically when storage was provisioned and exported to a host. Discovery of these components was a one-time effort for us.  Eliminating these time consuming and error prone processes had a dramatic effect on OPEX savings in our study, where ViPR Controller could save up to 100 percent, or $86,943, per year in administrative costs for SAN zoning alone.

Sure. ViPR Controller uses custom workflows to perform end-to-end provisioning operations which standardizes and simplifies the implementation process.  ViPR can also simplify the architecting and design steps through intelligent storage port selection. In ViPR 2.2, port allocation is based on performance metrics, computed metrics, and user-defined maximum limits with VMAX, VNX for Block, and Hitachi Data Systems (HDS).  ViPR collects the following storage port data points:

• Port percent busy
• CPU percent busy
• Number of initiators using a port
• Number of volumes using a port

Using these metrics, ViPR can avoid performance issues due to overloaded ports with too many volumes or too high of an I/O load.  Also, performance issues resulting from ports that reside on CPUs where the CPU percent busy is too high or the CPU is servicing too many volumes can be avoided.

ViPR takes averages of the performance-based metrics collected from the storage arrays and the number of initiators and volumes that it has already allocated to ports, and then compares these metrics to maximum limits (ceilings) configured by the administrator to determine which ports to select.

ViPR can automatically create zones and activate zonesets on Brocade and Cisco Fibre Channel switch fabrics.  Prior to zoning, ViPR must first determine the HBAs (initiators) and storage ports (targets) to use.  Through the process of discovery of the fabric (in the GUI, Physical Assets > Fabric Managers), ViPR creates an internal zone map.  It is aware of the which initiators and targets are connected and logged into the switch.  ViPR is also aware of the VSANs and Virtual Fabrics created by the switch administrator.

Once ViPR creates the volume, it performs an export operation.  Looking at the storage ports in the vArray and “SAN Multi-Path” attribute in the vPool, ViPR uses the intelligent port selection process described above to determine which storage ports will be used for the provisioning request.  The array configuration is completed first.  (As examples, the VMAX Masking View and HDS Host Group are created.)

Now, ViPR knows which initiator and target ports to use in a zone.  In version 2.2, ViPR Controller will reuse existing zones whenever possible.  This has the benefit of decreasing the load on the FC switch and reducing overall provisioning time.  If there isn’t an existing zone, ViPR creates one, adds it to the active zoneset, and finally activates the zoneset.  ViPR performs this operation on both A and B fabrics.

Thank you John and Vinay.

Are there other benefits of standardizing on an automated storage provisioning process?

Absolutely. We also considered the ongoing training costs associated with managing multiple storage platforms. Each storage administrator must have sufficient familiarity with every managed storage platform, which can add to expense.

We therefore examined the costs of vendor-specific courses designed to train storage administrators for platform-specific management. The courses we reviewed for NetApp, HP, Hitachi, and EMC averaged $4,237.50 per class. (The pricing details are in our report.) We then used this average as the cost of a generic course for any storage platform, and assumed that each employee attends at least one professional training course per year. At that rate, training a single storage administrator at each data center on only a single storage platform would cost $25,425. If a busy organization staffed each shift with two administrators to fulfill requests and then trained each administrator on all four storage platforms that we cited previously, training costs could skyrocket. Without ViPR Controller, training lower level storage administrators could cost a large enterprise organization over $610,200 (USD) per year.

ViPR Controller can reduce training expense by allowing administrators to provision storage without becoming familiar with any particular storage vendor’s proprietary method. ViPR Controller executed tasks based on repeatable, best practices, and intelligent processes. This means an administrator could provision storage from EMC VMAX and VNX storage using the same procedure used to provision storage from Hitachi, NetApp, or the other third-party storage providers.

Definitely. Many organizations with complex infrastructures, large customer bases, and contractual obligations to provide designated levels of service have turned to ITSM to provide a framework for enhancing the quality of management of their services. Change management is a big part of ITSM; its purpose is to minimize risk by carefully reviewing possible points for human error and to prevent outages by thoroughly understanding the effect a change will have on the rest of the environment and by taking any recommended mitigating steps.

One of the more expensive aspects of change management is the Change Advisory Board (CAB). The expense is purely a function of the amount of time required to review and approve changes and the staff necessary to make those decisions.

ViPR Controller can relieve the costs associated with a CAB by removing the work of a significant number of staff from the change phase. Again, it is all about mitigating the risk of humans introducing errors into the environment. By pre-allocating storage infrastructure, devices, disk pools, and other resources to ViPR Controller, organizations can avoid frequent infrastructure changes. Administrators, or possibly the users, fulfill requests in a timely manner without the introduction of management overhead, which can create a flexible organization with fewer human resources required for fulfilling requests.

In our study, we estimated ViPR Controller could deliver an 18% reduction in OPEX related to the CAB related costs, which for the example organization resulted in up to $1,319,733 over the course of a single year.

Did your study find other opportunities to reduce operating expenses?

Capacity planning/purchases represents a major investment for storage teams, did the study address savings in these areas?