Veeam Cloud Edition and Amazon Glacier Pricing

[rileybloke]

Hi,

I am about to trial Veeam Cloud Edition, and selecting which Cloud provider.
I read that Amazon Glacier is the cheapest, but I read on some forums and blogs about soaring retrieval costs.

I look after 2 companies, 1 SMB total VM size 300GB, veeam backup files are 120GB
And 1 SME total VM sizes 3TB, using replication not backup so I have yet be determined backup file size.

The SMB does not have offsite storage, and would like to have it, my SME use LT04 tape rotation, when compared to Veeam cloud approx £400 per year cheaper to run (excluding cloud service charges)
We need to backup offsite every night, and keep up to 10 restore points.

Our storage requirement is tiny compared to some of the pricing examples, which are quite hard to follow.
I understand Glacier costs maybe high for retrieval, but we would only need to retrieve data if the building collapsed, Veeam B&R is also used replicate/backup to DR hosts and NAS boxes on same site.
My concern is how Veeam manages the Amazon Glacier files/archives with an offsite backup policy like ours, deleting VBK or VIM files less than 3 months old would incur additional costs.
And does Veeam download any data from Amazon Glacier before the backup job starts, which may incur additional costs?

Has anyone used Veeam Cloud with Amazon Glacier to measure the costs yet?
Im interested and would be very greatfull if anyone could share any experience or information that may clear it up for me

Thanks
Riley

[dellock6]

Hi Riley,
I used Glacier for a couple of customers, and the use case was exactly what you described: SME customer, with no big budget to have an offsite storage to replicate to. As you stated, the price saving coming from Glacier is base on the low probability you would have to restore a file from it.
About pricing calculation, remember first we are talking about offsite backup, not replication (as you wrote). To get an idea, you can directly use Amazon calculator (never tested the one inside Veeam Cloud Edition yet):

http://calculator.s3.amazonaws.com/calc5.html

Retention policies of the Cloud Backup are completely separated and indipendent from VBR itself, you can keep bakup files into the cloud even once they are deleted from local backups due to expired retention (that is what long-term retention is for). When you need to restore, you need first to download the VBK file from the cloud provider, and from her start your restore activity.

Luca.

[zak2011]

I talked to an Amazon Galcier rep yesterday and honestly dont find it very convincing to go for this option due to some of the reasons.
Say your backup data to be copied to the cloud is 7TB.
1) Cost of storing the backup ofcourse is cheap (70$ per month to store 7TB). However to restore the data you pay 15000$ for 4hours RTO ( This is the time that Amazon promises your data will leave from their premises. Then you need to take into account how good your internet bandwidth is)
2)Let's say you have a 500Mbps fiber connection, it takes almost half a day to download 1TB which means you dont get much benefit of paying the 15000$. You might as well pay the 2500$ for 24hr Amazon's RTO.
Given the exhorbitant cost of data retrieval and very low RTO, perhaps it would be better to have external hard drives to store offsites.

The following article was a good read too.
http://www.technewsworld.com/story/72421.html

Thanks

[veremin]

In the way I see it, you might have misunderstood the Amazon Glacier concept; and it, in its turn, results in you being unsatisfied with it.

As Luca’s already mentioned, Amazon Glacier should be only considered as a long-term storage for data which stays unclaimed for long period of time. And it’s not something very interesting/affordable if you’re after high-availability of your information.

In contrast, if you’re eager to find cheap remote storage for saving your information, which is likely to be unclaimed for a long time, amazon glacier is exactly what you need.

Just my thoughts.
Thanks.

[zak2011]

Thats true. As long as every organization is very clear on the pros and cons, its always an individual choice.

[rileybloke]

Thanks everyone for the information.
I checked out the calculator for Glacier.

For my SME it would cost approx £3,000 to restore my data in 1 day.
For my SMB £200

My SMB is made up of 3 sites, the primary site has 25MB/10MB fibre, the other 2 sites have limited 2MB/512K ADSL due to geographical location. I have tried Veeam B&R backup over WAN links between the sites but it takes to long.
My options are removable tape/drive system, or cloud provider uploaded from the primary site with Fibre. I think I will go with Glacier for my SMB instead of investing in a tape solution.

As for my SME it is not feasible to ditch the tape library just yet, but at a time when we outgrow it, then we would probably shift to a cloud service(not glacier), in favour of ditching the same-site DR host(replica) and tape system(backup).

Thanks again
Riley

[Richard14]

Hi All,

I want to estimate the Glacier costs for my clients that will have 200GB to 1000GB's of data. They are all going to backup locally so I am interested in Glacier only if the building burns down, etc. I am not yet clear on how Amazon's policy on deleting objects younger than 3 months will affect my fees. I have pasted Amazon's official policy on early deletions below. I would likely want to do a reverse-incremental job to Glacier at least once a week but perhaps every night if bandwidth, etc permits. (I am thinking that the reverse incremental job would cause the least amount of data that is considered "deleted" but please correct me if I am wrong).

Can anyone shed some light on how Amazon will charge for these jobs in terms of early deletion of data? If there will be penalties on the deleted or overwritten data as the Veeam jobs run, would I come out ahead by storing the data for three months just to avoid the penalty? In other words, I would be storing more data total because I would be retaining it for longer. Would the early deletion fees be greater in my case than the additional storage cost? Are there any settings in Veeam Cloud Edition to help manage the retention to help in this situation? Specifically, could I configure Veeam to Not delete any data younger than three months AND would Amazon's policy then not charge for early deletions? Specifically, when the changes are applied to the full backup file, are those "changes" seen as deletions of the old data? If anyone has any experience on managing this I would be very appreciative! Also, has anyone had experiences with any of the other providers that Veeam supports in terms of pricing?

Thanks!

Richard

Official Policy:
"Amazon Glacier is designed for use cases where data is retained for months, years, or decades. Deleting data that is archived to Amazon Glacier is free if the objects being deleted have been archived in Amazon Glacier for three months or longer. If an object archived in Amazon Glacier is deleted or overwritten within three months of being archived then there will be an early deletion fee. This fee is prorated. If you delete 1GB of data 1 month after uploading it, you will be charged an early deletion fee for 2 months of Amazon Glacier storage. If you delete 1 GB after 2 months, you will be charged for 1 month of Amazon Glacier storage."

[foggy]

(I am thinking that the reverse incremental job would cause the least amount of data that is considered "deleted" but please correct me if I am wrong).

Richard, please keep in mind, that Veeam B&R Cloud Edition copies entire backup files, not changed blocks only. If you are using a reverse incremental mode on a daily basis, every day the entire vbk file will be uploaded and, hence, deleted according to retention.

Are there any settings in Veeam Cloud Edition to help manage the retention to help in this situation? Specifically, could I configure Veeam to Not delete any data younger than three months AND would Amazon's policy then not charge for early deletions?

Veeam B&R Cloud Edition has its own retention policy settings applied to the backups stored in the cloud storage allowing you to keep backups longer in the cloud than in your local repository.

[tsightler]

There is no way to do a "reverse incremental" job to Glacier, or really to any cloud storage with Cloud Edition. The job of Cloud Edition is simply to make it easy to copy the contents of your current, local repository to a cloud storage provider. Since a reverse incremental job running locally effectively creates a new full backup every night, you will the be sending an entire full backup to the cloud each night, or each week.

As far as I can tell, the "penalty" for early deletion of data from Glacier is equal to the costs had you simply left the data in place. In simple terms, while you can delete the data, your still going to be charged as if you had left the data in place for the minimum requirement. I'm sure this is to encourage people to use Glacier only for what it's intended for, i.e. long term archival.

Cloud Edition retention is separate from local backups so of course you can keep data in the cloud much longer than locally.

[Richard14]

Thanks Foggy and Tsightler.

The fact that Cloud Edition just copies the full backup files falls pretty short from what this should be in my opinion. Every company is going to put a premium on its relatively expensive internet bandwidth so only having the option to copy all your data over the wire won't fit well for most clients that I have. It is somewhat better than having a second utility manage the copies over the internet. It would be great if Veeam could partner with some cloud provider and give them a virtual appliance that will receive the backups so you could do reverse incrementals, deduplication, etc. and save your bandwidth. It sounds like I will be better off setting up a Windows VM in the cloud and install Veeam's backup proxy software on it to help receive the backups. It would have been nice to be able to use the glacier storage for its very low price point. If you all have any other thoughts I would love to hear them.

Thanks again!

Richard

[rileybloke]

I think the cloud edition is a brilliant idea for all SMB and SME that cant afford or have the skills of developing their own API's to interact with the Cloud providers. It provides the interface and control needed to utilise cloud services. Definitely a step in the right direction for Veeam.

Check out Twinstrata, they offer a free virtual appliance for Veeam customers.
And Rackspace provide cloud servers for which you can install Veeam proxy.

Food for thought

[dellock6]

Richard,
for your needs, why don't you install Veeam Proxy inside an EC2 machine? You can use it to do reverse incrementals in the cloud directly, and than send backups into S3 or Glacier from there.
I'm not getting the advantage for Veeam to eventually partner with 1 Cloud Provider, better than supporting many of them at the same time with the new Cloud edition. This gives you choice.

Luca.

[jcmachadouga]

I am with Luca, but this Amazon cloud being new to me, I have a million questions. I also came to the conclusion (I may be wrong) that (at least for our case) having a Veeam Proxy as an EC2 machine will work for us, but I don't really understand HOW I can send my backup files to Glacier from there.

Is anybody here already doing that?

Thanks a lot.

[seapro220]

I've looked into the 'cloud storage' over the last year or so and find, that for a SMB, it still isn't affordable. With me having a small amount of data, about 1.5Tb of data - not to mention long storage of that data (quartely and yearly) the costs for using 'a cloud' for such storage wasn't the best method. Instead I've decided to (cloud) store all my data onto a NAS box and maintain that data remotely. I have replications in place, and if i need to obtain last quarters backup - I can simply either download, or drive across down - download, and bring the data back within about an hour. Perhaps I'm fortunate being a SMB and can afford to undertake these measures, and in the future, as all the 'cloud storage' buzz settles down - the pricing will become more affordable for SMB's to rethink/retool their enviroments.

[topry]

We are a small business and have used Glacier since it's release and find it perfect for data that you must keep for long periods of time [7 to 10 years], and have another more easily accessible backup in the form of a removable hard drives stored in a data safe. We are in a heavily regulated industry and are required to follow 3-2-1-0 for some data, so Glacier serves as a low-cost copy (one of the 3) that we don't anticipate needing, but in case we must access it the retrieval costs would be less than any potential fines.

I would not use any cloud storage vendor for Veeam backups as I don't think it is well suited for that model. Even at our small size our vbk files we take offsite each week are over 2TB - a size that is impossible to try and push over our 100Mbps link.

We replicate mission critical VMs daily, the rest on the weekend or month end and once a week take a full copy offsite to a datasafe. Veeam is a great solution that has greatly simplified our lives, but the cloud option isn't viable for us. Now if we could get one of those Google gigabit links, I might give it try but alas, the infrastructure trolls in our area are fighting that initiative.

[dellock6]

First, a note to veryone complaining about Glacier costs for download and its policy to wait several hours before a download is available: Glacier is configured exactly to be e perfect fit for "long term" archive. IF one needs frequent restores, look to something else.

For incremental backups to Glacier, installing a repository role to an AWS virtual machine can do the trick: you ship only changed data to amazon once, inside the EC2 VM, and from there you can send a VBK to Glacier direclty from the Amazon network, at full speed. I'm completing a blog post, it would be hopefully ready in the next weeks.

Luca.

[zak2011]

Thanks Tim for sharing your experience in the real life. It was good to know.
Also thanks to Luca for sharing that tip. Will be looking forward to your post!

[dellock6]

Yeah, don't wait too much, I've a huge backlog of tests and blog posts

[cparker4486]

Yeah, don't wait too much, I've a huge backlog of tests and blog posts

Looking forward (a lot) to the post you mentioned above. No pressure though.

[lsmoryl]

I'm attempting to price out data backup costs to assess whether it's a better long-term investment to go with S3 or Glacier.

Aside from the obvious storage cost stuff (S3 is ~$85/TB*month, while Glacier is $10/TB*month), there are data transfer fees which are the same for both services. However, in addition to this Glacier has a data retrieval fee that scales almost exponentially with the inverse of retrieval speed--so a 4 hour retrieval for 1TB of data costs about $2000, while retrieving over a 24-hour period costs much less--only about $350. I'm curious whether it's possible to stagger data retrieval from Glacier by Veeam Cloud Backup in order to reduce costs (for my purposes, 24 hours is not too long to wait if it means over $1500 in savings).

Additionally, Glacier charges for premature (<90 day) deletions. It is unclear to me how that would play with Veeam Cloud Backup (for example, if I have 14 restore points on disk and am backing up each day, do I delete an image every 14 days?). If anybody has any experience/insight into the matter, I'd greatly appreciate it.

Many thanks,
Luke

[veremin]

Hi, Luke,

I'm attempting to price out data backup costs to assess whether it's a better long-term investment to go with S3 or Glacier.

As mentioned above, if you want to have long-term storage for the data that is likely to be unclaimed almost forever, then, chose, Glacier. Otherwise, S3 might be a better option due to less price for data retrieval.

It is unclear to me how that would play with Veeam Cloud Backup (for example, if I have 14 restore points on disk and am backing up each day, do I delete an image every 14 days?)

In order to copy data to the chosen cloud, you should create backup plan first. The number of restore points kept in the cloud is controlled by "purging" options of backup plan. These settings are not affected by the retention settings of the "source" backup job.

Thanks.