The following is a recording and full transcript from the webinar, “How To Reduce Public Cloud Storage Costs”. You can download the full slide deck on Slideshare.
Full Transcript: Consolidate Your Files in the Cloud
John Bedrick: Hey! Welcome to SoftNAS webinar on how to reduce public cloud storage costs. We’ll be starting officially in about one to two minutes to allow all the attendees the option of getting in at the beginning of the webinar. Thank you, and just stay tuned.
Hey! We’ll get started in about one minute.
John: All right, welcome to SoftNAS webinar on how to reduce public cloud storage costs. My name is John Bedrick. I am the head of product marketing management at SoftNAS, and I’ll be conducting this webinar.
Please use the chat capability on the side to ask any questions which we will address during the question and answer session at the end of our webinar.
If you wish to reach out to me after the webinar, to ask some specific questions that you didn’t want to ask in a public forum, please feel free to do so. My email address is firstname.lastname@example.org. Let’s get started.
What we’re going to cover today are just a few items. We are going to cover the trends in data growth. We’re also going to cover, sort of as a primary if you will, what to look for in public cloud solutions.
From there, we’re going to transition over to how SoftNAS can help with our product SoftNAS Cloud 4. We also conduct a brief demo of our brand-new Storage Cost-Savings Calculator which will help you understand how much you can save on utilizing SoftNAS Cloud 4 in a public cloud storage environment.
Last, we’ll have some closure remarks, and then we’ll follow it up with the Q&A. Again, please use the chat facility to ask any of your questions.
The amount of data that has been created by businesses is staggering – it’s on an order of doubling every 18 months. This is really an unsustainable long-term issue when we see what IT budgets are growing at compared to businesses.
IT budgets on average are growing maybe about 2 to 3% annually. Obviously, according to IDC, by 2020 which is not that far off, 80% of all corporate data growth is going to be unstructured — that’s your emails, PDF, Word Documents, images, etc. — while only about 10% is going to come in the form of structured data like databases, for example. And that could be SQL databases, NoSQL, XML, JSON, etc.
Meanwhile, by 2020, we’re going to be reaching 163 Zettabytes worth of data at a pretty rapid rate. If you compel that by some brand new sources of data that we hadn’t really dealt with much in the past, it’s really going to be challenging for businesses to try to control and manage that when you add in things like Internet of Things, big data analytics, all of which will create gaps between where the data is produced versus where it’s going to be consumed, analyzed, backed-up.
Really, if you look at things even from a consumer standpoint, almost everything we buy these days generates data that needs to be stored, controlled, and analyzed – from your smart home appliances, refrigerators, heating, and cooling, to the watch that you wear on your wrist, and other smart applications and devices.
If you look at 2020, the number of people that will actually be connected will reach an all-time high of four billion and that’s quite a bit. We’re going to have over 25 million apps. We are going to have over 25 billion embedded and intelligent systems, and we’re going to reach 50 trillion gigabytes of data – staggering.
On the meantime, the data isn’t confined merely anymore to traditional data centers so the gap between where it’s stored and where it’s consumed and the preference for data storage is not going to be your traditional data center anymore.
Businesses are really going to be in a need of a multi-cloud strategy for controlling and managing this growing amount of data.
If you look at it, 80% of IT organizations will be committed to hybrid architectures and this is according to IDC. In another study from the “Voice of Enterprise” by the 451 Research Group, it was found that 60% of companies will actually be upgrading in a multi-cloud environment by the end of this year.
What to do. While data is being created faster and the rate of the IT budget is growing, you can see from the slide that there’s a huge gap, which leads to frustration from the IT organization.
Let’s transition to how do we address and solve some of these huge data monsters that are just gobbling up data as fast as it could be produced and creating a huge need for storage.
What do we look for in a public cloud solution to address this problem? Well, some of these have been around for a little while.
Data storage compression. Now, for those of you who haven’t been around the industry for very long, data storage compression basically removes the whitespace between and in data for more efficient storage.
If you compress the data that you’re storing, then you can get a net benefit of savings in your storage space and that, of course, immediately translates into cost savings. Now, all of this cost is subject to the types of data that you are storing.
Not all cloud solutions, by the way, include the ability to compress data. One example that comes to mind is a very well-promoted cloud platform vendor’s offering that doesn’t offer compression. Of course, I am speaking about Amazon’s Elastic File System or EFS for short.
An EFS does not offer compression. That means; you either need to have a third-party compression utility to compress your data prior to storing it in the cloud on EFS or solutions like EFS, and that can lead to all sorts of potential issues down the road. Or you need to store your data in an uncompressed format; and of course, if you do that, you’re paying unnecessarily more money for that cloud storage.
Another technology is referred to as deduplication. What is deduplication? Deduplication sounds and is exactly what it sounds like; it is the elimination or reduction of data redundancies.
If you look at all of the gigabytes, terabytes, petabytes that you might have of data, there is going to be some level of duplication. Sometimes it’s a question of multiple people may be even storing the exact same identical file on a system that gets backed-up into the cloud. All of that is going to take up additional space.
If you’re able to deduplicate that data that you’re storing, you can achieve some significant storage-space savings, translation into cost savings, and that, of course, is subject to the amount of repetitive data being stored.
Just like I mentioned previously with compression, not all solutions in the cloud include the ability to deduplicate data. Just as in the previous example that I had mentioned about Amazon’s EFS, EFS also does not include native deduplication.
Either, again, you’re going to need a third-party dedupe utility prior to storing it in EFS or some other similar solution, or you’re going to need to store all your data in an un-deduped format on the cloud. That means you’re, of course, going to be paying more money than you need to.
Let’s just take a look at an example of two different types of storage at a high level. What you’ll take away from this slide, I hope, is that you will see that object storage is going to be much more cost-effective especially in the cloud.
Just a word of note. All the prices that I am displaying in this table are coming from the respective cloud platform vendors on the West Coast pricing. They offer different prices based on different locations and regions. In this table, I am using West Coast pricing.
What you will see is that the more high-performing public cloud block storage costs are relatively more expensive than the lower performing public cloud object storage.
In this example, you can see ratios of five or six or seven to one where object storage is less expensive than block storage. In the past, typically what people would use that object storage for would be to put less active storage and data into the object storage.
Sort of more of a longer curve strategy. You can think of it as maybe akin to the more legacy type drives that are still being used today.
Of course, what people would do will be putting their more active data in block storage.
If you follow that and you’re able to make use of object storage in a way that’s easy for your applications and your users to obtain too, then that works out great.
If you can’t…Most solutions out in the market today are unable to utilize access to cloud-native object storage and so they need something in between to be able to get the benefit of that.
Similarly, being able to get cloud-native access to block storage also would require access to some solution for that and there are a few out in the market and, of course, SoftNAS is one of those. I’ll go into that a little bit later.
If you’re able to make use of object storage, what are some of the cool things you can do to save more money besides using object storage just by itself?
A lot of applications require high availability. High availability is exactly what it sounds like. It is maintaining a maximum amount of up-time for applications and access to data.
There has been an ability to have two computing instances access a single storage pool — they both share access to the same storage pool –in the past, on legacy on-premises storage systems and it hasn’t been fully brought over into the public cloud until recently.
If you’re able to do this as this diagram shows — having two computer instances access an object-storage storage pool — that means you’re relying on the robust nature of public cloud object storage. The SLAs typically for public cloud object storage are at least 10 or more 9s of uptime. That would be 99.999999% or better, of up-time, which is pretty good.
The reason why you would have two computer instances is that the SLAs for the compute is not the same SLAs for the storage. You can have your compute instance go down in the cloud just like you could on an on-premises system; but at least your storage would remain up, using object storage.
If you have two compute instances running in the cloud, if one of those — we’ll call it the primary node — was to fail, then the rollover or fell over would be to the second compute instance, or as I’m referring to on this diagram as the secondary node, and it would pick up.
There would be some amount of delay switching from the primary to the secondary. That will be a gap if you are actively writing to the storage during that period of time but then you would pick back up in a period of time — we’ll call it less than five minutes, for example — which is certainly better than being down for the complete duration until the public cloud vendor gets everything back up.
Just remember that not every vendor offers this solution, but it can greatly reduce your overall public cloud storage cost by half. If you don’t need to have twice the storage for a fully high-available system and you can do it all with an object storage in just two compute instances, you’re going to save roughly 50% of what the cost would normally be.
The next area of savings is one that a lot of people don’t necessarily think about when they are thinking about savings in the cloud and that is your network connection and how to optimize that high-speed connection to get your data moved from one point to another.
Traditional ways of filling lots of physical hard drives or storage systems, then putting them on a truck and having that truck drive over to your cloud provider of choice. Then taking those storage devices and physically transferring the data from those storage devices into the cloud or possibly mounting it can be, one, very expensive and filled with lots of hidden costs. Plus, you really do have to recognize that you run the risk of your data getting out of sync between the originating source in your data center and the ultimate cloud destination, all of which can cost you money.
Another option, of course, is I’ll lease some high-speed network connections between my data center or my data source and the cloud provider of your choice. That also could be very expensive. Needing a 1G network connection or a 10G network connection, those are pricy.
Having the data transfer taking longer than it needs to means that you have to keep paying for those leased lines, those high-speed network connections, longer than you would normally want.
The last option which would be transferring your data over slower more noisy error-prone network connections, especially in some parts of the world, is going to take longer due to the quality of your network connection and the inherent nature of the TCP/IP protocol.
If it needs to have a retransmit of that data, sometimes because of those error condition or drops or noise or latency, the process is going to become unreliable.
Sometimes the whole process of data transfer has to start over right from the beginning so all of the previous time is lost and you start from the beginning. All of that can result in a lot of time-consuming effort which is going to wind up costing your business money. All of those factors should also be considered.
The next option I’m going to talk to you about is one that’s interesting. That is, assuming that you can make use of both object storage and block storage and be able to use them together. Creating tiers of storage where you are making use of the high-speed higher performing block storage on one tier and then also using other tiers which would be less performance and less expensive.
If you can have multiple tiers, where your most active data is only contained within the most expensive higher performing tier, then you are able to save money if you can move the data from tier to tier.
A lot of solutions out in the market today are doing this via a manual process. Meaning that a person, typically somebody in IT, would be looking at the age of the data and moving it from one storage type to another storage type, to another storage type.
If you have the ability to create aging policies that can move the data from one tier to another tier, to another tier and back again, as it’s being requested, that can also save you considerable money in two ways.
One way is, of course, you’re only storing and using the data on the tier of storage that is appropriate at the given time, so you’re saving money on your cloud storage. Also, if it could be automated, you’re saving money on the labor that would have to manually move the data from one tier to another tier. It can all be policy-driven so you’ll save money on the labor for that.
These are all areas in which you should consider looking at to help reduce your overall public cloud storage expense.
I’m going to transition, here now, to what SoftNAS can offer in helping you save money in the public cloud. We are going to address all of the same areas that we talked about before plus a little bit more besides.
The first one that we had started with before was compression. SoftNAS cloud contains a compression utility – it’s built into the product, it’s no extra charge, it’s just one of our features.
SoftNAS utilizes LZ4 lossless data compression algorithm that’s focused on compression and decompression speeds so they are optimized for speeds.
Of course, your storage space savings is going to be dependent on data type. However, on average, your savings will fall somewhere between 2X and 90X of space savings. Yes, that’s quite a range.
If you look at the table that I have off to the right, just as some examples, you can see the percent range for some common file types. You can see that, for example, a text file can go from 0% all the way up to 99% space savings with an average space savings of about 73%.
Then you can see some Word files, Executable files, and Photo files. You can see things like picture files in a jpeg format. They’ll compress down as well as some other files but that’s going to vary.
Again, just to give you some idea that data type does have an absolute impact on ultimately the compression ratio that you are going to achieve and the space savings that that’s going to translate into.
If there’s any other questions on compression a little bit later, we can pick them up during the Q&A.
The next is deduplication. SoftNAS Cloud is built upon Open ZFS. Our in-line deduplication is passed over to us via ZFS in-line dedupe. The storage space savings from deduplication, again, is totally dependent upon your data redundancy. However, on average, it’s going to fall somewhere between 2X to 20X of space savings and that’s also going to translate into a tremendous amount of storage space savings and significant cost-savings in the public cloud.
Now, I realized I broke these two apart which I did also at the beginning of the webinar. But if you combine the two, both the compression from LZ4 lossless compression and the ZFS in-line deduplication, you could see that there can be very significant amounts of savings in your public cloud storage.
Again, just a reminder. Not all public cloud offerings contain deduplication and compression. Again, as an example, EFS does not. If you have 1TB of data that’s uncompressed and not deduplicated, if you put it into EFS, you’re going to require 1TB of storage so you gain nothing from the fact that your data could have been deduped and could have been compressed. Just keep that in mind as you’re looking at public cloud offerings.
The next thing I want to talk to you about is SoftNAS ObjFast and that’s our object storage accelerator. The ObjFast capability within SoftNAS Cloud makes object cloud storage like, for example, AWS, S3 or Azure Blob as an example storage. It helps it run very close or near to block level performance while taking advantage of object storage pricing, which can result in cost-savings on storage of up to two-thirds. You could save up to two-thirds of your storage costs by using SoftNAS Cloud and our feature for this called ObjFast.
How does ObjFast work? I’ve realized that we have a lot of bar charts here. By the way, this chart is for Azure. I have another one I’ll follow up with on AWS.
SoftNAS ObjFast makes cloud object storage like Azure Blob run at near block-level performance. You can see that you can actually achieve savings of up to 70% of your cost-savings versus using block storage alone. How does ObjFast work?
ObjFast works by throttling data transfer to cloud object storage so we achieve a speed that is as fast as possible but without exceeding Azure object storage read/write capabilities. I’ll talk a little bit more about that in a bit, but let’s switch over to Amazon – AWS.
One other point that I wanted to make. All of the charts that you see for Azure are using the DS15 compute capability. All of the graphs that you see on this slide in Amazon are using an Amazon EC2 instance of c5.9xlarge, just to put that into perspective.
Again here, ObjFast is making cloud object storage like AWS – S3 run at near block-level storage and you can get resulting savings up to 67% on Amazon versus block-level storage alone.
Next. I’m going to go into what SoftNAS is calling our Dual Controller High Availability or DCHA for short. DCHA uses object storage only and it’s an excellent way of leveraging the object storage resiliency to lower costs as well as lower replication overhead – meaning you don’t need to have two destinations of data.
The public cloud vendors, they’ve worked really hard to make their object storage offerings highly scalable and reliable, over (10-12) 9s of uptime. That means that you can achieve high-availability using a single storage pool of object storage versus the need to double that storage amount as required by regular high-availability as done by SoftNAS and some other vendors that use replication to move from one of the storage pools to another.
Using SoftNAS’s patent-pending Dual Controller High Availability, you can save at least half of your storage that you would need for standard high availability, and of course, that’s going to result in significant public storage cloud savings.
Switching to the WAN optimization section that we had talked about earlier, SoftNAS has a patented UltraFast technology and this saves costs by saving on the time needed to accelerate global bulk data.
UltraFast is able to operate up to 20 times faster than standard TCP/IP and also at one-tenth of the cost of alternative bulk data transfer solutions. SoftNAS UltraFast accelerates the transfer of data into, out of, and across private/public hybrid clouds and multi-clouds.
UltraFast can overcome up to 4,000 milliseconds of latency and up to 10% of packet loss due to network congestion connecting remote locations with the cloud over any kind of network conditions. This is also a significant time-saver as well as a money saver.
The next feature I want to talk about that’s within SoftNAS Cloud is our SoftNAS SmartTiers. SoftNAS SmartTiers is our patent-pending automated storage tiering feature that moves aging data from more expensive high-performance block storage to less expensive object storage, and this is all done according to your own customer-set policies while reducing public cloud storage cost.
What makes it interesting — the SoftNAS solution — is you could see in this slide that there is multiple tiers of storage. Let’s call tier 1 a high-speed block storage, tier 2 might be a medium speed block storage, and tier 3 might be a less performing object storage.
From an application or a user-perspective, SoftNAS hides all that and to the application and to the user, it looks just like any other cloud volume. You would access it just like you would any other drive or network share, but behind the scenes, there’s these multiple tiers of storage that are all policy-driven for moving data up and down from tier to tier, to tier and back up again.
In the case of SoftNAS SmartTiers, we’re saying that you can save up to 67% on AWS or up to 70% on Azure of your public cloud storage cost. We do this by moving data blocks from tier 1…in this chat, up to tier 3. Although, in our user-interface, we can support up to four tiers if you really need that.
We move the data blocks from tier 1, through tier 2, to tier 3, and even tier 4 if you’ve configured that via a policy that you set up. And then we will migrate it back up if you need access to that data.
Obviously, the goal is to only keep the most active data in the most expensive storage types and move the data as quickly as possible to the least expensive least performing tier of storage.
Let’s take a look at an example. This is AWS. If I had 100TB of storage that I needed to store and if I were to put it all into the high-performing block-level EBS storage at Amazon SSD, my cost would be about $123,000.
If I was to put it all in object storage assuming that I could, based on the application, then my cost would be about $37,000. That’s quite a jump between object storage and block storage.
By using SmartTiers and if I chose a ratio of 20% in block-level storage and 80% in object level storage. I would put 20TB in EBS and 80T in S3 and my cost would be about $54,000, which is fairly significant in terms of what the cost of EBS would be.
Let’s take a look at the same thing but in Azure. In Azure, I could save up to 70% using the same example. If I had 100TB in Premium SSD on Azure, that would cost me about $142,000. If I was to do the same 100TB in object storage in Azure Cool Blob, it would cost me about $19,000.
Then again, if I was using SmartTiers and I did the 20%-80% ratio, where 20% is in premium storage and 80% is in object storage, then I would be at $43,000 rate, so fairly significant.
We’re nearing the end. I’m going to quickly show you a demo of getting to the Smart Tiers cost-savings calculator for cloud storage.
You navigate to the SoftNAS web page. Then you would scroll down and you would see a “Discover How Much You Could Save On Public Cloud Storage Cost” and a “Calculate My Savings” button. You click on that and you get taken over to the SoftNAS SmartTiers Storage Cost Savings Calculator.
This is meant to be as easy as possible. You come over to “Choose a cloud platform vendor.” In this example, I’m going to choose Amazon Web Services, AWS. The UI wants to know how much you’re going to put of data initially. I’m going to put in 50TB.
Will I be using high-availability? In this example, I’m just going to say no. How many tiers of storage will I be using? I’m going to say three.
I’m going to choose how much year-over-year data growth am I going to assume I’m going to have. On this calculator, you can go up to 500% data growth. But in my example here, I’m doing 20%
What’s my initial storage type? Within Amazon, I’m going to choose one of this high-performing block-level storage so I’m choosing Provisioned IOPS. In my tier 2, I’m going to take Medium Magnetic. For my tier 3 storage, I’m going to go ahead and pick S3 Standard Storage.
How much do I want to keep in tier 1 and how much do I want to keep in tier 2 as a ratio? Those defaults are good for me. The tier 3 by default would be 70%.
I get a bar chart automatically generated for me. It says that my three-year cost without SmartTiers is the bar on the left, and my three-year cost with Smart Tiers is the bar on the right. Also, it calculates what my savings would be over three years.
If I really want to get a detailed report, all I need to do is fill out this form. This is live information. Yes, that’s my real name. I type that in. I type in my company. My company email again, email@example.com if you need to reach me.
Once I fill in all that information, I’m just going to click the “Download my detailed report” and automatically, you’re going to see your report will be generated for you. It’s thinking and thinking. Eventually, in the lower left…
I am using Chrome, in this example, for my browser, by the way. You’ll see the report. I’m going to click on that but you can save it. What I would advise you to do is save your report to your local system.
This is what your report will look like. As we scroll down, we’re going to confirm your entries for your selections. I chose AWS. For my tier 1, 2, and 3, those are the types of storage that I chose.
Scrolling down, you’re going to see the same familiar bar chart that you had for the 3b and all the numbers are there. You now need to hover to see what those are. A savings of almost $215,000 over three years, using Smart Tiers in this example.
What’s my storage growth? Because my data in tier 1 was chosen to be 10%, tier 2 was 20%, and tier 3 by default is 70%, you can see the growth of data over three years in each of those tiers.
What’s my savings over three years? In each year, you can see what my cost would be without SmartTiers and what my cost would be with Smart Tiers. And what my savings would be, in these green bar charts, in year one, two, and three, by using Smart Tiers.
We also give you a table with the actual numbers. You’ll be able to plug those in to any analysis that you’re doing in selecting cloud vendors. That’s all available to you. We hide nothing.
Then we also give you breakdown pie charts for three years. Year one; between tier 1, 2, and 3. Year two; your cost between tier 1, 2, and 3. And year three; your cost between years 1, 2, and 3.
Please visit the website to get to the calculator at softnas.com. Don’t forget you can also pick up the telephone and reach any of our sales people as well.
Why do customers choose SoftNAS? We have a number of very selective customers who need to have their data available to maximum uptime.
The first reason that they chose us is that we’ve been in business since 2013 and we’ve been cloud specialists all that time. We’re cloud native experts. Cloud is what we live, eat, breath, sleep day in and day out.
I also invite you to reach out to any of our cloud experts for consultations so that we can help you with your journey to the cloud. We help you with cost-savings without sacrificing performance.
A lot of people could say that they can save you some money and some may and some may not. If you add in that second part, “Without sacrificing performance,” that’s a critical aspect.
I showed you those bar charts of our performance, of accelerating object storage for example. We are also doing fast in our block level storage access as well, and pretty much, we will be one of the fastest offerings out there in the market.
Our support team is second to none. We have received more accolades on our support team than we could spend during this webinar talking about.
One of the last couple of things that I want to mention is that we’re very flexible in our offerings so that you can pretty much adopt it and grow with it as your business grows, whether that’s on-premises to private cloud, on-premises to public cloud in some kind of hybrid option or even multi-cloud scenarios.
We have all of that flexibility and our technology just works. That’s what we hear from our customers – it just works.
In terms of some of our customers that trust us — and these are all international companies for the most part — you could see that we have pretty much of who’s who in this space and they run all different industry verticals. Cross industry verticals running from SMB up to the Fortune50 all trust SoftNAS with their data.
At this point, I am going to pause and see if there are any questions that I can answer for you as we go along during this webinar.
We have one question that came in. The question is we say that we’re less expensive than Amazon Elastic File System or EFS but we didn’t account for the costs for the compute instance which is included with EFS. That is correct.
We do actually have a slide that I can jump to that will show you an example of EFS. Here, you can see the same slide that I had before except, now, I’m using EFS.
You can see, with EFS, it’s about 369,000 for that initial terabyte and the S3 and the Smart Tiers is still the same. I didn’t build into this slide what the compute costs would be for EFS.
If I click and reveal the next slide, you’ll see that I just picked three EC2 instances and you can see how far up the costs of those, for three years, would be. For those EC2 instances added in, you can see we’re still significantly cheaper than Amazon EFS. Hopefully that answered your question. Looking for the second question.
The next question. Is your ObjFast patented? Our ObjFast technology is patent-pending and we do expect to receive a patent on that. I do want to mention that SmartTiers is also patented-pending. Our UltraFast technology is patented and not just patent-pending.
Looking around, do we have any other questions from anybody? No. All right. I appreciate everybody’s time here today. We thank you. we hope that you’ve gotten a little bit of some ideas of how you can save costs in the public cloud.
We invite you to do several things. First, if you want to learn more, go to our website at softnas.com. We also offer a 30-day free trial so please feel free to go to softnas.com/trynow.
If you are wanting to speak to one of our cloud experts, again, go to softnas.com. There should be a chat feature available on our website. Feel free to go ahead and connect with that and you will be connected with one of our cloud experts.
Last but not least, again, if you would like to reach out to me directly to clarify or answer any other questions that you felt shy about asking in a public forum, please feel to do so at firstname.lastname@example.org.
With that, we are going to wrap up this webinar. I thank you again for your attendance. It was a great pleasure to be able to present this to you. Thank you, and have a great rest of your day.