Podcast: Ruthlessly prioritise applications during Covid-19

In this podcast, we talk to Pure Storage CTO Alex McMullan about the effects of the Covid-19 lockdown on organisations’ IT requirements.

McMullan talks about how the transition to remote and 24/7 working requires a scaling up of storage and compute or a possible re-prioritisation of the resources that applications are provided with.

He also discusses how organisations can move to utility-based consumption models of IT that can be scaled up, and scaled down more easily when we arrive at the “new normal”.

Download this podcast

Antony Adshead: What are the impacts on storage of a rapid scaling up of remote working?

Alex McMullan: In regard to storage, I think there are some obvious challenges we’ve already been dealing with, but also some that are a little further down the line that we need to think about.

So, if you think about storage scaling up to support everyone working at home remotely, I think that’s fairly obvious in terms of what that means.

It puts more pressure on all the traditional infrastructures. Things like the VDI side of things – the broker – but also if you think about how that impacts particularly in larger infrastructures where there’s a consequential demand for everybody who is impacted, let’s call it software-defined networking or network function virtualisation.

That’s an additional load that goes onto infrastructure such as firewalls or into WAN connections or VPNs or basic data protection that comes with that. So it’s not just the VDI impact, it’s everyone else working remotely on the infrastructure that’s, frankly, less well tested and less well used on a normal day-to-day basis.

What it also means, obviously, is that we switch very quickly from a 9-5 model where everyone went to work at the same time, but now many of us are parents, we’re looking at workload sharing, whether that is looking after children, home schooling, cooking, shopping, but also those of us who are volunteering with the NHS as well.

We very quickly switch from that 9-5 model to 24/7. What that also then means is that it brings challenges into other parts of IT because many companies [are experiencing] a change freeze or are accelerating deployments of new infrastructure or expanded capacity or simply doing software updates. And there’s a bilateral impact here in that we’re seeing more demand for change, but also more demand for continuous and always-on services to go with that.

And that’s also driving a bifurcation in the way that cultures are dealing with that. We’re seeing [parts of] organisations working together, verticals such as the storage team, the network team, the virtualisation team, the compute teams had been dealing with their own expansions at different times in different ways. Suddenly everyone in every business is asking for more of everything at the same time.

What that also then drives is the forecast for the second wave, which is obviously the way in which we’re living our lives for the next few weeks.

We’re seeing more demand on online retailing; food delivery, of course, but also the second order consequences of that, such as contact centres, our banks, our car insurance, mobile phone providers. They’re all seeing a huge uptick.

Obviously, videoconferencing, too, is driving more demand for storage, but also because of the desire to record more of these events for a podcast or a webinar or simply for more wide distribution within companies. Everybody is consuming more recording capability in that videoconferencing and we see that very clearly in our own customer base.

The final part is that there is obviously more and more pressure coming from the e-commerce side.

Again, more billing, more transactions, more online payments, but also there are second order consequences in terms of fraud detection and prevention because the biggest part of the remote working challenge that we see – which is partly a storage problem and partly not – in everyone using shared infrastructure at home.

Not every family has five PCs, [one] for every member of the family, but they are sharing, so are using the same infrastructure as our five-year-olds, our teenagers and that’s obviously giving the bad actors a lot more, a lot wider attack surface to play with because of that shared infrastructure on PCs, laptops and tablets.

So, lots of AV, lots of malware detection, lots of data loss prevention, lots of intrusion detection happening, not just inside datacentre edges, but outside the company periphery. All of these are putting more stress on software as we transition to that almost BYOD-at-home capability.

So, lots of things to think about there. Broadly, everybody wants more of the same, and more quickly. And I think we’ll talk later on in the podcast about ways of addressing, absorbing and perhaps solving for the future in some of these trying times.

Adshead: What can organisations do to scale up storage capacity against such a broad range of impacts during this period?

McMullan: I think it depends on the size of the organisation. There are different things that can be done depending on whether you’re an enterprise or a small business.

If we take the enterprise first, the key thing there is that enterprises today should have CMDB (configuration management database), an application register, and a BC/DR plan.

At this point in time, every business should understand which applications are critical, which are nice to have and which are effectively “if there’s nothing else to do”. So, if you have two datacentres, one for disaster recovery, you should be considering which of those lower criticality applications can be shut down on the primary site and only run in disaster recovery, ie their disaster recovery beyond that is a restore from tape backup or archive backup or disk-based backup.

It’s only the highest criticality applications that should maintain that BC/DR and they can then start to consume production infrastructure capacity that was being used for those lower-tier applications.

The key part of any strategy that involves this is to communicate effectively to business owners and application owners that you’re doing that because it is something that they’ll have to deal with over time.

That’s the first thing in terms of reusing what you have today. The second thing is you can ask your users to free up capacity, but my own experience working at some of the biggest banks in the world tells me that you won’t get much benefit from that. If everybody deletes 0.1% of their storage capacity, you don’t really get any benefit from that. It’s always worth asking, but my own experience has been that you don’t get a huge benefit from it.

The third part of that, again for enterprises, is to talk to your cloud providers, service providers and managed service providers about utility-based options you can consume from them, whether it’s dollars per gig per month utility or simply short-term capacity. We’re even seeing customers asking for short-term loans, and that’s something we will look at on a case-by-case basis.

So, there’s three ways there of not just having to make large capital investments but also to get that on a utility-based basis and, of course, if you’re using AWS or Azure or Google or Rackspace, any of the public cloud or hosting utility providers, you can always do that, although it will depend by region which capacity is available at any given time.

If we then switch and think about the smaller businesses that perhaps don’t have that same cash-flow beneficial pool to draw on that the larger businesses do, I think again they can look at talking to managed service providers or the utility providers or the public cloud providers for short-term burst capacity, and I think that’s vital for them. But again, they should be looking to move lower-criticality applications into those use cases and keep what’s most important for them on-premise.

The other thing that comes with that is that we shouldn’t think about this being a short-term thing, and I think we’ll come back to that in a moment, but I would say in answer to that for everyone is, take the opportunity now to revisit what you’ve got, what is import and what applications and what datasets are useful to you, and if you do have applications that aren’t as business-critical, perhaps now is the time to consider retiring those and decommissioning them in general.

Adshead: How might organisations reverse the kinds of changes they’ve made if/when things go back to normal?

McMullan: I think the answer there is two-fold – partly technical, partly cultural. So, let’s deal with the technical side first.

I think, in terms of returning to normal, based on what I’ve seen over the last few weeks, that the way that we work is going to change going forward, full stop. I think there have been some benefits from the rapid innovation cycles and the ways that people have been communicating more effectively and I think we will carry some of that culture going forward as we switch back to the “new normal”.

I think that will end up being some way between the way we used to work and the way that we are currently working.

The investments that companies make in utility storage, utility compute and managed service hosting, those can obviously be reversed and terminated when the contract ends, so that’s all fine.

And you can also then spend time reversing the changes that larger enterprises have made in terms of restoring the business continuity and DR capabilities of the lower-criticality applications and bringing them back up to the full recoverability point.

The plan then for that next phase, as it were, gets much more interesting. I’m not an epidemiologist, I’m not a biologist, but to me as an engineer, it’s fairly clear that to a point where a sufficient part of the public is vaccinated and we have a good set of testing, we should expect further waves, whether seasonal, whether in six months’ time, but I think we will be back in that same place again, perhaps further down the road.

So the key thing here is that we carry forward the ability to rapidly innovate, the ability to change quickly and adopt a similar can-do attitude as we have seen from all parts of the community this time. Because there will be a next time and we’ll be a lot better prepared next time. The key thing is not to forget how we were easily able to adopt and change our postures when we had to. That, for me, is the key thing to take away in terms of scaling down and returning to normal.