One of the articles I wanted to write for some time now is about the decision process of a backup infrastructure replacement. Unfortunately for me and my fellow backup consultants, most of the customers see a backup infrastructure as an enormous hidden cost with no “added-value” and therefore it has a lower priority in the budget lists. On these moments, you need a qualified IT team member or backup consultant to motivate why we need those kind of budgets. Not an easy task if I can say so!
I always tell them… don’t see your backup environment as a cost, but more as an insurance! In the end, aren’t we all paying our fire insurance and hoping we will never need it?!? In the field, I see 2 kind of persons. Persons confronted with data loss and persons who will be confronted with data loss. So it’s about time to raise some awareness!
For security investments you can find various Return On Security Investments (ROSI) calculators on the internet (http://www.rositech.com/ for example). These will help you understand what you are paying is nothing compared to what you can lose after an intrusion, an exploit or virus outbreak. So I’m wondering why did nobody started with a Return on Backup Investments (ROBI) calculator? The financial calculation of these risks are beyond the scope of this article and to be honest I think they need to be addressed by a non-technical guy or squad (represented by the CFO, key business owners, a representative of the IT team and – if needed – an external consultant).
This article is intended for IT managers who are responsible for the technical part of the decision-taking process. I need to disappoint people expecting a lot of numbers, financial calculations and analyses of different solutions. Unfortunately I’m a bit to technical to do these calculations and I’m convinced every customer environment differs from each other.
My personal moto is:
“Build an end-to-end solution which covers your current needs and provides the flexibility to change and grow with the technology and future business requirements!”
I strongly believe the choice of a new backup solution needs to be addressed with the highest amount of precaution, analysis and validation. It’s not a storage, nor a server replacement where after a 4-year cycle you just get everything new! The main reason is it interacts with basically everything and choosing a new backup software will still require the old environment (software & tape media) to remain online to ensure historical restores. These requirement can cause a lot of headaches and should be avoided by executing a good and thorough investigation and product analysis.
We can conclude we will not change our backup software every 4 years when we have permanent retention in place. But what is permanent? 25 to 50 years depending from industry to industry? I simply don’t know it and I’m quite sure nobody does.
This brings me to the first lock-in of the backup solution.. The software! Choosing the right software for today and tomorrow is key in the decision process! During this article we will touch some other lock-ins as well.
First of all I want to talk about the buzzword of 2013/2014: cloud adoption!
Don’t be mistaken, the cloud is here to stay! Everything will differ on how the cloud will be addressed from within the backup solution! I strongly believe we will see a trend returning again “consolidation – silo approach – consolidation” and now we are going back to a “silo approach”.
Originally we started from a mainframe and we moved it to standalone servers. This approach gave us no extended availability of the platform and did cost us a lot of money on hardware and maintenance. So we started virtualizing everything again and look at it in every direction, everybody accepted it and has a virtualize-first approach. And… above all each phase had its own backup approach!
Today we notice customers are starting to adopt the cloud. Bit by bit they are moving away from their on premise solutions and are storing corporate data in the cloud! These type of environments are called ‘hybrid cloud environments’.
This trend performs a re-allocation of the budget from CAPEX (investments) to OPEX (operational costs). From a technical point of view, we lose the flexibility because it’s ‘just a service’. On the other hand from a managers perspective, we did a handover of the responsibility and continue to work with the cloud provider with a pre-defined service level agreement. Additionally, we are able to do a more accurate cost estimation (for example 10€ per user/month).
However from a backup perspective we will notice the silo’s returning again:
- In case e-mails are required to be recovered, use the Office 365 solution.
- In case user data needs to be recovered, recover it from the recycle bin in your OneDrive for Business.
- In case system data needs to be recovered within your Azure or Amazon environment, use the cloud backup solution.
- In case you still need to recover something from an on premise solution, use the on premise solution.
Additionally I don’t think enterprise environments are ready to move entirely to Platform-As-A-Service (PAAS) provider. I’m convinced they will adopt it in different phases (test/sandbox, DR, hybrid & adopt) over a certain time period. For the system integrators out-there.. It will take one to maybe two infrastructure refreshments cycles before cloud migrations will be daily routine. So start preparing yourself on a technical level, but as well on a legal level (IT governance)! Consultancy will be key in making the difference! Mark my words.
One of the next steps I expect to see in enterprise backup environments is the transition of the archives to a cloud storage container. Archives tend to require slow IO requirements as they are not actively used. Pushing this type of ‘static data’ will remove the need of recurrent backup protection.
In a next phase we will see production systems being pushed into the cloud as a hot stand-by in case of disaster. One step later, we will start running production in the cloud.
The game changer within the backup vendor industry will be the one adopting the cloud by providing a single-pane-of-glass to reduce complexity and ease management. I’m quite sure there will be some technical innovations at stake the coming years.
The three-ilities: recoverability, flexability and portability!
One of the things I put as a must-have in choosing the right solution is recoverability, flexibility and portability. I don’t want to use a solution where I need to restore to the same platform by using traditional agent-based backup/restores!
- I want to have the flexibility to virtualize my server when it crashed.
- I want to have the recoverability to perform bare-metal recovery in case of a system crash.
- I want to have the portability to burst out to the cloud if I don’t have the necessary storage space or computing resources available for recovering systems.
- I want to have the ability to backup systems from remote sites and replicate my backup data to a central location for on-the-fly disaster recovery compliancy and remove the need for site-local tape handling. This centralization optimizes the site-local investments and operational costs.
- I want to have replication mechanism between on premise and cloud providers resulting in portability.
- I want to have the ability to perform storage based snapshot backup to provide a hot-standby backup methodology and archive to tape functionalities. And I would like to have this storage-vendor-independent!
- I want to have archiving functionalities for the most common platforms (mail, collaboration, file, …) and I want to store it only once!
- I want to have end-user search for user-driven granular recovery.
I simply want an end-to-end solution providing a single-pane-of-glass which provides me transparency and ease of management. It’s just that easy!
That’s why I would advise… look at it from a higher level!
First of all, in my line of work I tend to avoid the word backup. I rather refer to ‘data management‘, because that’s what we do as a ‘backup consultant’.. we manage your data, we secure it, we provide end-user recoverability and provide data insight with build-in archiving and searchability for end-user search or legal compliancy!
Don’t look at your backup environment from a silo approach, because you will hit the boundaries at a certain moment. Try to define the relationships with storage, virtualization, cloud, applications, end-user data protection/recovery, remote sites, disaster recovery and your future plans (‘the IT Infrastructure Roadmap’). These relationships are the foundation of the business- and user requirements of the new backup solution!
Take a closer look at your data and try to categorize it in tiers regarding importance and level of protection. Once again, don’t look at it as a silo but include also other areas to ensure the best solution.
|Tier||Description||Recovery Point Of Objective|
|Platinum||Clustered system with synchronous replication with daily snapshot offload||Zero|
|Gold||Clustered system with asynchronous replication with daily snapshot offload||Near-to-zero|
|Silver||Standalone system with 4-hourly incremental and daily backup||Worst-case 4 hours|
|Bronze||Standalone system with daily backup||Worst-case 24 hours|
I did not mention ‘Recovery Time Objective’ in the table above, as it’s pretty much impossible to calculate it in the Silver and Bronze tiers. Let me explain why..
- We don’t know the size of the system! In case we need to do a full system restore, it’s obvious 200GB will be more swiftly restored than 2000GB.
- We don’t know the configuration (virtual machine, multiplexing on backup-side, multi-streaming, …)
From a pricing point-of-view we can illustrate the price as follows. By implementing data management (archiving in this case), you can reduce the price of the front-end system by moving data to “archives on-premise or in the cloud”. There is no harm in thinking outside the box!
So what’s next?
After categorizing data in different tiers and identifying the relationships, it’s time to go on the market and reach out to different suppliers/vendors for a solution and a quote. My first step would be consulting the Gartner quadrants for:
- Backup & recovery;
- Cloud adoption (storing data in the cloud & securing data in the cloud);
- Storage based snapshots backup & recovery.
These will give you already a good impression about the leaders on the market. Based on your ‘IT infrastructure roadmap’ it’s now time to create an Request For Information (RFI), or – if you have the required info already available – an Request For Proposal (RFP). Don’t forget to ask the suppliers roadmap regarding new business trends, cloud adoption and technological innovation.
I would advise to reach out to multiple vendors even though you already ‘made your choice’. There is no harm in collecting some additional information and who-knows maybe you missed something. Additionally it’s an opener during the discussion about price.
The roots of the backup environment!
As previously mentioned, a backup software is something which correlates with each component holding data in your environment. Ensure that the design does not create any new lock-ins such as:
- Storage based snapshot offloading where the destination needs to be the same storage array or a newer version;
- The inability to do media refreshments, resulting in old tape devices which need to stay in the environment for historical restores;
- Backup appliances offloading data to tape creating a lock-in from the backup appliance.
Additionally regarding the backup appliances and/or regular disk storage. It’s important to keep in mind, these need to be phased out before the contract ends minus the maximum retention on disk (which will cause an elevated CAPEX). Another way is to copy the data to a new solution, but these operations can take a significant time to complete. So I’m just wondering.. depending on your retention.. is the squeeze worth the juice?
Thanks for reading,