SoftwareDefinedChat

Too much of the budget going toward backup. Need to cut cost in order to shift more budget toward growing the business

Many legacy backup systems don’t keep up with the times. Virtual environments, for example, have changed the way data needs to be protected. Cloud is another…

>> Too much skill and administrator time being spent on operating complex backup systems

they seem to lose control of their infrastructure and just react to problems of the day

Growing amounts of data are making it increasingly more difficult to manage backups

Scale. Data growth certainly isn’t slowing down. In many cases it has simply outgrown old approaches and tools for backup. Can no longer hold the data, fit in the backup window or meet the recovery objectives.

Denis Vo, is it scale that causes them to lose control?

they focus on cost, which leads to commodity/inferior solutions, which leads to scalability, so I'd say yes to scale

I find that many are dissatisfied bc decisions on backup and recovery are made as after thought not primary

but flash has changed that trend so it's interesting to see how new faster infrastructure changes the game. Love to get thoughts on that

@JohnFurrier Say more on how Flash is impacting backup dissatisfaction...

@RonRiffe IT guys I talk to say Flash changed the conversation bc of how the data is moving and stored so that forces the back up an recovery converation upfront

@furrier good point.

Have seen Flash being used for metadata for backup, but actual data goes on other storage technologies. Anyone seen other uses of flash in backup?

Dissatisfaction abounds in other areas also, where demands on employees went way up in the economic downturn (do more with less), and demands continue to outpace investment in people. A tough time for IT folks having cloud mandated on them.

@SwiftMario mandates, lower budgets, do more with less... How do you cope?

I think more than coping, models are breaking down, and IT folks are somewhat forced into finding different ways to do what needs to be done. LOB folks are choosing SaaS apps instead of using internal IT, and IT is using more cloud to cope.

you have to look at the ratio of compute to capacity. This is very difficult to size - and in dev led agile environments, impossible to capacity plan. So scale out-data orientated flexible infrastructure is essential - that's Spectrum!

many organisations have yet to map services to apps and apps to actual servers. Classifying the data to be protected via a Data Protection Catalogue allows backup and recovery platform to be designed in accordance with business need

Well, first, IBM has been spending a LOT of money developing a portfolio of software defined storage offerings. There were some household names in there like SAN Volume Controller and Tivoli Storage Manager.

But earlier this year we went about rebranding and expanding that portfolio. We introduced IBM Spectrum Storage. Here’s a blog link that explains the portfolio http://ow.ly/NI8p0

That said, our offering that does data protection is IBM Spectrum Protect (formerly Tivoli Storage Manager)

Spectrum Protect has been focusing on a few key areas that our clients have been telling us and the analysts are projecting...

Things like: 1.) hybrid cloud and being able to efficiently protect your distributed datacenters and remote back offices with different storage capacities and retention requirements

Things like: 2.) Protecting virtual environments with powerful function to backup your entire datacenter very efficiently with CBT, but still enable instantaneous restore and access at the granularity you need.

full capabilities interacting with the available hypervisor APIs - and extra value like instant, thin recovery of VMs - effectively managing the copy data problem at the virtualisation layer

Things like: 3.) Snapshot support with Spectrum Protect Snapshot and the ability to leverage the hardware snapshot features in both IBM and non-IBM storage devices to have application integrated protection with complete federation between applications.

also remember that multi tenant virtualisation is rapidly evolving into application orientated container environments, some based on openstack and @docker - which we have great capabilities for.

and: 4.) Software-based deduplication, included in the Spectrum Protect shrink-wrapped box. Store your data fast and efficiently regardless of what hardware you have, fully compatible with cloud object stores too!

And also 5.) simplicity. The fact is we all have too many things to do during the work day. We are working with the best UX designers in the business to make sure we build EASY into everything we do, from initial discovery, to deployment, to day to day use

@RonRiffe I read the blog a couple of days ago and it´s very straight-forward and easy to understand what IBM Spectrum Storage is bringing to the industry

great question SLAs are changing too with apps driving the new metrics or KPIs - it's wild west at this point so whoever comes to the table with cloud and on prem can do well

Good question... I think it depends on how you use cloud.

Sending backup copies directly to cloud has implications... backup speeds, recovery times...

I think @RonRiffe said it well. There is still a need for some onsite backup to achieve quick recovery, but cloud can be used for longer term retention of backups

There are techniques our customers are using that leverage hybrid configurations

What is helpful is to have a backup solution that can help you manage multiple methods of backup to multiple tiers of storage depending on SLAs, including a Cloud tier

Most recent or most critical copies onsite and older, less critical archives in the cloud

@RonRiffe - it's the recovery that I'm asking about. Glacier, as an example, charges for recovery - and takes (up to) days. By then, an enterprise is dead. So - how does one create a rapid recovery environment, yet leverage the cost structure of cloud?

the recovery capability needs to be an integrity point for the business service. This needs to traverse on-prem and cloud systems that provide the service. Analytics are the bridge to the could ensure the data has the right recovery capability

ie we have the intelligence in real time to use cloud for the right data at the right time. and to ensure data born in the cloud is protected in line with the service level.

@iantsmith I completely agree, analytics are critical in the growing complexity of hybrid cloud environments

right! @Werner_Sam a service level has to be policed - or it is immediately irrelevant if it cannot be trusted.. i.e. do you really know that you can recover a service in the time you have set in the policy? Real time analytics sure this.

Disk was always faster than tape, and dedup was first about saving cost. While not as efficient on stored bits, private cloud storage can be just as dense and lower in cost.

With hybrid cloud, local restores are just as faster (actually faster in parallel across a cluster), and consistency to offsite locations does the vault-like offsite protection. One way to design for SLAs, keep more onsite for fast restore.

@SwiftMario I've always like parallelisation of recovery! like a torrent swarm! get the data back in the most efficient manner...

I go back to hybrid, the bulk of backup copies in the cloud on some ultra cheap service, the critical copies on premises... hybrid lets you balance cost and performance

certainly seeing a focus on disk (efficiently used) for backup specifically, and acceptance that tape is well suited to archive, ideally in the cloud and well away dark backup data

Do you see other data protection strategies being employed in cloud other than backup? For example, multiple copies and redundancy without an offline version being kept?

Yes, for most web applications, there is no offline version, objects are distributed as uniquely as possible across a cluster, which can span multiple geographies, where the 2nd site is an online but redundant copy of data.

IT managers are starting to think of the cloud in the same way as they thought about multiple of their own datacenters. The same use cases apply for cross-site replication

We find that even with replication uses cases, backup copies and archive copies still play a critical business role

There is a good paper on DRaaS that includes other cloud backup solutions that you might like: https://www.ibm.com/...