Cloud or local

Forums › LAN (ethernet/Wifi/routers/Bluetooth) › Cloud or local

Yorkshirekid

193 posts

Master Geek

#288851 27-Jul-2021 18:54

Not sure if this is the correct forum, sorry.

I'm looking for some arguments FOR keeping data in the office instead of in the cloud.

Assuming we have adequate hardware and software security, what would be some solid arguments not to go cloud based?

I’m aware of the advantages of using the cloud, but looking for positives for local storage.

Thanks all

1 | 2

BDFL - Memuneh
79309 posts

Uber Geek

Administrator

ID Verified

Trusted

Geekzone

Lifetime subscriber

#2750900 27-Jul-2021 19:05

If it's one office only then local storage will likely be faster than cloud. Synology for example offers good features for local data backup and you can then backup the NAS data to the cloud, creating a hybrid model - fast for local users but still secure with off-site backup.

1024kb

1168 posts

Uber Geek

ID Verified

Lifetime subscriber

#2750905 27-Jul-2021 19:28

The ransomware attacks are coming through Cloud services. What a great attack vector too! Hack one company, infect all their clients. Remember, another name for Cloud is “somebody else’s computer”. That’s the most compelling reason for local data retention.

In amongst the one million infections that the rEvil Kaseya attack achieved were 500 Swedish Coop supermarkets & 11 schools in New Zealand - yup, encrypted files right here in our own back yard. If you rely on cloud 100%, then you are exposing yourself to the strong possibility of a recurrence.

Your 3rd sentence made my skin crawl. “Assuming” and “adequate” are two words that do not belong in this conversation. Assuming will lead to overlooked shortcomings, adequate will be inadequate within short timeframe.

There’s solid arguments for both sides of this argument. The real best solution is hybrid - go both. If your cloud provider should fail, you have a local copy. & another off-site local copy. Robust solution would be 2 geographically separate Cloud services alongside 2 local storage copies - one on-site but using tape, not HDD, the other can be HDD but stored off-site.

Yes, that’s a lot of redundancy. It’s also a lot of surety. If it saves your ass just once, you’ll be in profit.

Megabyte - so geek it megahertz

timmmay

20589 posts

Uber Geek

Trusted

Lifetime subscriber

#2750918 27-Jul-2021 20:25

Performance and latency would be better. It can be cheaper to have your own equipment, depending on what you're doing.

If you do go on-premise make sure you have really good, incremental, offsite backups, and that the backups are regularly tested.

ANglEAUT

2327 posts

Uber Geek

Trusted

Lifetime subscriber

#2750978 27-Jul-2021 23:08

timmmay: ... make sure you have really good, incremental, offsite backups, and that the backups are regularly tested.

Thx to the Kaseya incident, many are saying that even with having good & tested backups, restoring that amount of data in one go would take to long. Restoring a file here or an instance there is doable, but not when you have Petabytes of data that needs to be restored a.s.a.p.

Please keep this GZ community vibrant by contributing in a constructive & respectful manner.

michaelmurfy

meow
13265 posts

Uber Geek

Moderator

ID Verified

Trusted

Lifetime subscriber

#2750987 28-Jul-2021 00:10

I'm just a user with a home lab setup but my data is stored locally (SSD cached storage) that is uploaded each night, with versioning to Crashplan. I've restored from Crashplan multiple times when I've screwed up and deleted something I wasn't supposed to and it has been great. I've also pulled larger disk images off Crashplan with great success.

I wouldn't go 100% cloud based (with the exception of email - which should be looked after by the major companies for email delivery reasons) but anything local, ensure whatever it is running on is supported, patched with a password policy in place especially if you're dealing with potentially confidential data. We've seen ransomware attacks (eg - the DHB attack) where confidential data was released online once the attackers found they were not getting paid - it doesn't matter that you've restored from backups already by this stage, your confidential stuff could have forced cloud adoption.

Michael Murphy | https://murfy.nz
Referral Links: Quic Broadband (use R122101E7CV7Q for free setup)

Are you happy with what you get from Geekzone? Please consider supporting us by subscribing.
Opinions are my own and not the views of my employer.

freitasm

BDFL - Memuneh
79309 posts

Uber Geek

Administrator

ID Verified

Trusted

Geekzone

Lifetime subscriber

#2750994 28-Jul-2021 06:44

1024kb: The ransomware attacks are coming through Cloud services. What a great attack vector too! Hack one company, infect all their clients. Remember, another name for Cloud is “somebody else’s computer”. That’s the most compelling reason for local data retention.

"Cloud" involves so many different services - software, platform, storage, identity, security. The incidents you are referring to are mostly when people use some software or another. They are not related to storage directly (although never say never) and providers like Microsoft Azure, Backblaze B2, AWS S3, HPE Cloud Storage are completely different from "someone's colocated computer being shared."

Dynamic

3869 posts

Uber Geek

ID Verified

Trusted

Lifetime subscriber

#2751005 28-Jul-2021 07:26

Do a quick cost analysis in Excel based on a 3 year period. If I do X solution, what do its setup and ongoing costs come to over 36 months and what are the benefits of doing it this way. How does this compare if I do Y?

If you use Microsoft 365 services, you may be able to use OneDrive/SharePoint at no extra monthly cost.

Plan for the worst possible data loss scenario. We adopted a client recently who (as the company owner) is an awesome guy to deal with and is a little techy. He was very proud of his NAS setup and that he had it replicating to another NAS for data recovery. This had been in place for 18 months before we met them. They wanted to be a casual client, having us help our rather than taking over management of their network. That was fine with us, but there are some minumum requirements that we have. It didn't take much of a discussion for him to approve our managing a backup for them.

Less than 6 months later their files were crypto'd via the QNAP NAS zero day attack, and of course the encrypted files replicated to his second NAS. We restored their files from our managed backup and migrated them to Microsoft 365 storage pretty quickly. After that, we required the backup of the NAS files and added backup of the files stored in the Microsoft 365 system.

“Don't believe anything you read on the net. Except this. Well, including this, I suppose.” Douglas Adams

Referral links to services I use, really like, and may be rewarded if you sign up:
PocketSmith for budgeting and personal finance management. A great Kiwi company.

Sharesies

Trade NZ and US shares and funds with Sharesies (affiliate link).

freitasm

BDFL - Memuneh
79309 posts

Uber Geek

Administrator

ID Verified

Trusted

Geekzone

Lifetime subscriber

#2751007 28-Jul-2021 07:31

Yep, people who think replication is the same as backup...

timmmay

20589 posts

Uber Geek

Trusted

Lifetime subscriber

#2751018 28-Jul-2021 07:58

ANglEAUT:

timmmay: ... make sure you have really good, incremental, offsite backups, and that the backups are regularly tested.

Thx to the Kaseya incident, many are saying that even with having good & tested backups, restoring that amount of data in one go would take to long. Restoring a file here or an instance there is doable, but not when you have Petabytes of data that needs to be restored a.s.a.p.

True. Personally I have backups on a disk near my house I run every month or so, plus daily cloud backups. I would restore locally then get the latest files from the cloud. Though actually since my data volume is low I'd probably go 100% cloud.

Restore time is definitely something to consider for people with a lot of data.

rogercruse

644 posts

Ultimate Geek

ID Verified

Lifetime subscriber

#2751037 28-Jul-2021 08:52

I have Microsoft 365 account - office software, 1 terabyte of cloud storage (photos, documents, whatever), security, etc.

If only they would improve their face recognition inside their Photos app (it isn't as good as Google Photos or Amazon Photos)

freitasm

BDFL - Memuneh
79309 posts

Uber Geek

Administrator

ID Verified

Trusted

Geekzone

Lifetime subscriber

#2751040 28-Jul-2021 08:57

rogercruse:

I have Microsoft 365 account - office software, 1 terabyte of cloud storage (photos, documents, whatever), security, etc.

But that isn't a backup.

gbwelly

1243 posts

Uber Geek

#2751046 28-Jul-2021 09:05

If you are talking 800TB->1PB type sizes you can easily build your own geo-redundant storage including decent snapshot retention and a backup solution for significantly less than cloud over a 5 year period. But if the organisation is high opex/low capex then it can be a hard sell.

ShinyChrome

1575 posts

Uber Geek

ID Verified

Trusted

#2751049 28-Jul-2021 09:13

Probably old news to the more experienced Ops hands here, but I just recently learned of the 3-2-1-1-0 strategy, which would seem appropriate to the argument for "why not both?"

Might not be appropriate for your volume of storage/data compliance, but my plan for my personal data was to have copies of files on devices cloud-synced via OneDrive or Google Drive to my on-prem NAS, which is in-turn backed up daily to Backblaze, Crashplan etc, as well as a local external HDD weekly. Ideally the HDD would then be stored off-site, but unless banks bring back safety deposit boxes, I don't see that happening.

timmmay

20589 posts

Uber Geek

Trusted

Lifetime subscriber

#2751059 28-Jul-2021 09:25

Raid is not a backup. A copy is not a backup. Incremental, offsite, immutable, multiple locations is the way to go :)

ghettomaster

387 posts

Ultimate Geek

#2751069 28-Jul-2021 09:45

Not wanting to railroad the discussion, but the Kaseya hack is actually a good argument FOR cloud services.

All the SaaS Kaseya clients were safe because as soon as Kaseya noticed something was up, their playbook was to shut down all their hosted Kaseya servers and they did just that. For all the on-prem customers, Kaseya had to reach out to each one and get them to shut down. Even though Kaseya did the right thing by reaching out as pro-actively as they could, such a process was always going to be too slow.

1 | 2