5 tips for a more efficient data center

5 Tips for Optimizing the Data CenterAdam RobertsChief Hardware ArchitectOctober 2015

SanDisk Confidential 2

This presentation contains certain forward-looking statements.

Any statement that refers to expectations, projections or other characterizations of future events or circumstances is a forward-looking statement, including those relating to market growth for flash and SSD, industry trends including the increased adoption of and demand for flash and SSD in various devices, enterprise applications and data centers, future memory technology, technology transitions and future products, including 3D NAND and its performance, cost and timing of commercial availability. This presentation also contains information from third parties, which reflect their projections as of the date of issuance of those statements.

Actual results may differ materially from those expressed in these forward-looking statements due to factors detailed under the caption "Risk Factors" and elsewhere in the documents we file from time to time with the SEC, including our annual and quarterly reports.

We undertake no obligation to update these forward-looking statements, which speak only as of the date hereof or the date of issuance by a third party, as the case may be.

Forward Looking Statement


Computer Engineering and Electrical Engineering Degrees from North Carolina State University. Starting with a 6 year career in Pharmaceutical Engineering, went on to hold various development and strategy positions in the Enterprise Storage Industry for the past 13 years. 5 of those years was the Head Engineer for IBM’s System x RAID and Storage Group followed by 7 years as IBM’s System x Flash Solutions Architect. Adam has over 120 patents issued or pending.

Adam RobertsChief Solutions Architect, SanDisk

Speaker Overview


Storage and compute together in a single unit

Lots of units operating in a cluster or “Cloud”

There are lots of ways to build a solution

or


What do these examples have in common?

It doesn’t matter the configuration, a change in one aspect of the solution

can affect the final capability of the entire solution


What do these examples have in common?

Example:It doesn’t matter how fast the CPU is if

the storage can’t provide data fast enough

It doesn’t matter how fast the storage is if the CPU can’t process the data

It doesn’t matter the configuration, a change in one aspect of the solution

can affect the final capability of the entire solution

Balance the variables to get the best result!


Tip #1 - Understand your workload

• Don’t build a solution based on wrong workload assumptions. (Too fast or too

slow)

• Don’t let lack of workload understanding cause you to build with excessive

performance margin

• Understand the balance of performance and write endurance.


Tip #2 - Understand the “access density” needed and which solutions and devices can provide it


Performance margin of SSDs allow for smooth RAID recovery

• HDDs usually have no margin for rebuilding arrays in critical mode

• No performance to spare so application suffers during rebuilds

• SSD performance margin can provide needed application performance while

rebuilding a critical array transparently in parallel

• Artificially boosting HDD access density by short stroking or using small capacity

devices cause wasted capacity and/or larger than necessary HW footprint


Tip #3 - HW consolidation is your friend!

Low cost NAND devices with large capacity points can allow:

• Smaller solution HW footprint

• Lower power consumption

• Reduced “Solution cost per GB”

• Better performance

• Fewer points of failure


*$669 for 300 GB 15k SAS HDD - $6,999 list price for 3.84 TB Optimus Max - $4,000 list price for DAS enclosure **Assumes 9 watt HDD - 7 watt SSD – 100 watt enclosure

Storage solution cost can be greatly reduced!


4. Don’t get lost in the components of the data center. What matters is the end result.

This allows for a complete reimagining of the data center!


Solution cost and specifications are the focus as opposed to single device specifications

• Convert a complicated circuit, or in our case a complicated solution, into simple basic needs of the datacenter

• Model solutions that can illustrate what can be done. Open our minds to what’s possible.

• Lets do things we have never been able to do before

Think Thevenin or Norton equivalent circuits…….

Summarize of what’s needed for the solution and design for it! Don’t design around functionally limited building blocks.


With the same performance can we operate more efficiently?


Tip #5 - Analyze and identify changes to help you fill them more efficiently

• Don’t simply build “the same as last time”.

• Question everything. Cost, capacity, level of protection, and performance, affect what you build.

• Example: A RAID 10 HDD solution may be used because of HDDs performance limitations. SSD performance margins may allow for the user to move to a capacity efficient RAID 5 solution.

• An application may have no use for SSD levels of performance but can benefit greatly from it during failure scenarios.


Performance margin of SSDs vs HDD allow for RAID configuration changes

• RAID array types in servers are frequently dictated by either capacity or performance sensitivity

• RAID 10 for performance sensitivity

• RAID 5 for capacity sensitivity

• RAID 6 to limit exposure in critical mode

– Ex. If forced to use RAID 10 to get HDD performance, SSDs can provide similar performance with RAID 5 or RAID 6, thus saving usable capacity

Performance versus Capacity

Vs


Enclosure count can be greatly reduced!


19


Other efficiency tips for the Data Center • In high growth areas where resources needed per server may change

over time we can benefit by resource pooling

• Remove the need to build half empty chassis

• Add resources as needed

• Remove the situation where you have to replace an out of date server but the storage/DRAM or CPU components are still good

• HW breakdown is time consuming

• Ex. Not enough CPU capabilities but flash devices in the box are still great


Questions to ask yourself

• Do you understand why your current architecture is configured the way it is?

• Do you really know your current workload and access density needs?

• Do you know the amount of capital expense tied to redundancy and oversubscribed HW, like small capacity drives or short stroking?


© 2015-2016 Western Digital Corporation or its affiliates. All rights reserved. SanDisk and the SanDisk logo are trademarks of Western Digital Corporation or its affiliates, registered in the U.S. and other countries. Other brand names mentioned herein are for identification purposes only and may be the trademark(s) of their respective holder(s).

Thank You!Questions?

[email protected]

@SanDiskDataCtr

@BigDataFlash itblog.sandisk.comSanDisk Data Center Solutions

5 tips for a more efficient data center

Technology