TEC Talk: COVID-19 & Impact on Office 365 Services

Welcome to our TEC Talk today sponsored by Quest Software. Today's TEC Talk is on COVID-19 and its impact on Office 365 services and how to enable your remote workforce. So right now, you're either in the fire of enabling your suddenly remote workforce, or you're already over that initial hump. You flipped on a bunch of Office 365 services, and now you're flying fast, making sure people can use them effectively and securely. That's why we wanted to offer something else of value, hearing directly from a Microsoft MVP, Most Valuable Professional, The Experts Conference Speaker and Petri.com contributor, Tony Redmond. He's going to be talking about the impact COVID-19 is having on Microsoft's ability to deliver their services in the face of a 775% increase in demand and outline the management implications of moving and combining chat, voice, video, and file sharing into a single platform. TEC Talks are not Quest product pitches. They're pure training in the spirit of our Active Directory in Office 365 Training Conference, called The Experts Conference, or TEC. The TEC Talk is just like our TEC conference-- training you can use with no third party pitches. So what is The Experts Conference? In 2019, TEC returned after a seven-year hiatus, and it was an amazing sold out success. This year we're taking TEC 2020 to the next level. So along with our Hybrid Active Directory security and Office 365 tracks, we're adding a third track called Migration and Modernization. We also carved out time for networking and interaction with those industry experts and your peers, folks like Tony Redmond, today's speaker, Sean Metcalf, Microsoft Certified Master, Windows Security Expert, Randy Franklin Smith, Microsoft Teams Product Manager Chris McNulty, David Kennedy, founder of TrustedSec, who's testified before Congress on national security and appeared on several national news and TV shows. We've got over 30 speakers ranging from Microsoft employees, Microsoft MVPs, to independent industry experts, so you're sure to get the technical insights on Microsoft technologies you're looking for. TEC 2020 is slated for November 17th through the 18th at the Loews Hotel in Midtown Atlanta, Georgia. We are optimistic that COVID-19 will be contained before TEC 2020, but we also understand that this is a new experience for everyone. We're planning for every possibility, and therefore, we are also offering a full refund to anyone who requests it by November 3, 2020. This way, you can register now, take advantage of registration discounts, while ensuring a full refund, should any changes on your part need to happen. We're also adding a pre-conference day for Quest product training around the Quest Active Directory Recovery and Quest Active Directory Auditing solutions and specific recovery and auditing scenarios. So while TEC itself has no third party pitches, there is a pre-conference day that is very much dedicated to those specific Quest lines. So if you are looking at those products or you own those already, this is certainly a place to come and get hands-on training, scenario-based training. So to learn more about the Experts Conference, please visit theexpertsconference.com, and register today for the early bird pricing before TEC 2020 sells out. Now on to today's speaker, Tony Redmond runs his own consulting company. He is a lead author for the Office 365 IT Pros e-book series. He writes twice weekly for Petri.com, covering all things Office 365. He is a frequent speaker at conferences, including our own Experts Conference, as you see here in the screen. And we're really excited that he is here to share his experience and what he's seeing on the impact that the surge in Office 365 usage is having on Microsoft's ability to deliver those services. So without further ado, I'm going to turn it over to Tony to dive into today's topic. And we will take Q and A at the end. So Tony, welcome. Thank you, Jennifer. Thank you. But I think the most important thing about MVPs that I ever heard in terms of defining what they are is that they're called independent experts. We're not there to champion Microsoft. We give Microsoft, actually, a hard time a lot of the time about the deficiencies that are in their technology. We try and work out on a very productive level, a proactive level to-- how they can improve matters. While I've got people here, let me start off by just giving you one big piece of advice, as you scale up Office 365 to [INAUDIBLE] dealing with right now. That is take a look at conditional access policies, and if you're not using them, make sure that you start considering at least using conditional access policies to make sure to keep your tenants safe. And please, please, please, enable multi-factor authentication. Microsoft will tell you that 99.9% of the basic attacks that are made against Office 365, especially against Exchange Online, are deflected or stopped dead in the water by accounts having MFA enabled. So right now, today, there's only about 12% of Office 365 accounts with MFA enabled. It's a shame. It's a disgrace. It shouldn't happen. So if you take anything away from this session, please go and enable MFA. So with that, let's get on. So I've got two sets of slides, or two-- the slides are divided in to two parts. First part's just talking about what's happening inside Office 365 now at the moment, what kind of stresses and strains are on the service, and how you might see that. And in the second part, we'll talk a little bit about how Office 365 is developing as it approaches the second decade of an Office cloud service. So please put in your questions, and we'll get down to them later on. OK, so let's move on. Technology is, once again, escaping me as slides are refusing to move. Oh, no, here we go. All right, so everybody on this call knows that we're all living under a huge pandemic at the moment, which is driving an enormous amount of need to work from home. We're getting it because the governments are telling us to stay at home. Here in Ireland, where I am right now, we're not allowed to move more than 2 kilometers from our house. But the good thing is that because the internet has improved dramatically over the last number of years-- by that, I mean the whole internet service, both the backbone service-- and because we now have a very robust cloud services infrastructure, people can actually do it. And I mean the point is-- for example, Microsoft, over the last number of years, has deployed literally thousands of point to presence connectors to allow people to make a very close connection to the Microsoft backbone. And once data gets into the backbone, it gets routed very, very quickly to all those various data centers that make up Office 365. If you look at the actual usage patterns, Microsoft came out last week and said 775% increase in cloud services generally now. That is spread across all of their cloud properties, and a lot of it was in Azure. But the one that was really interesting for us from an Office 365 perspective is what happened with Teams. Teams had been growing quite nicely. It's had a really good growth curve over the last while. But, you know, if you were to look back at Teams two weeks ago, you'd say it was about 30 million users, which, to put into perspective, is really only 15% of the Office 365 install base. And then bang, work at home took off. Now all of a sudden, Teams increased by 37 and 1/2 percent in a week. So you now get this thing that's 44 million daily active users connected to Teams. Now daily active user means somebody does something during the day. It's not just a matter of logging onto Teams. It's a matter of [INAUDIBLE] to schedule a meeting now, to read a message. You have to join a meeting. They have to take some sort of action that they have to explicitly make rather than implicitly perform, as in the case of a log on. Why Teams are so interesting is because Teams is built on the top of so many other parts of Office 365. So the growth in Teams automatically drove a huge growth in Exchange, in SharePoint, in OneDrive, in Planner, and all the other bits that get connected into a Teams ecosystem. SharePoint is probably the heaviest, because, of course, every team that's created comes with a SharePoint site and all of that sort of stuff, the [INAUDIBLE]. Another interesting factor is that we've seen some companies [INAUDIBLE] up increases of 100%. In fact, there's now a 3,081% growth in the number of meetings held over a one-week period. And that kind of increase on that infrastructure doesn't only just stress the cloud infrastructure that you're connected to, but it also impacts your own network, because clearly, data is not to get in and out of the cloud. Microsoft says there's 900 million minutes daily of meetings and calls. That actually isn't all that much per active user, but it's increasing all the time. In fact, it's increasing so much that Microsoft has accelerated plans to introduce a new codec to make a calling and signalling experience a lot better. So in that area, we have seen some increase as well in SharePoint, OneDrive, and also Stream. Stream is kind of interesting, because a lot of these meetings that are happening, like corporate meetings, corporate meetings tend to be recorded. Once they're recorded in Teams, there's a bot that actually hands off meeting to stream, and once it gets to the stream, the recording is processed. So it's processed for all the captioning. It's also processed so that different versions of the video is created for playback on different devices-- you know, if you were to play back something on a candy phone, on a phone, candy bar shaped phone, or an iPhone or whatever, you get a different experience than if you played back a Stream video on the PC or a Mac. So the interesting thing here is that when you, as a tenant, are set up for Stream, you get a storage allocation of 500 gigabyte, plus 1/2 a gigabyte per active user, per licensed user. And if you are recording Stream videos for Teams meetings, you're going to run out of that allocation quite quickly if you don't keep your eye on it. Microsoft, at one point, gave a guidance of about 2 and 1/2 megabyte per minute of recording. And, in fact, with Teams meeting recordings, it's turning out to be more like 7 and 1/2 gigabytes. So just keep an eye on that for your own benefit. So the big-- with all of this increase in demand, they're scared that the service is not going to be running quite as smoothly as it did. The interesting thing is that things have worked better than you might have predicted, and that's testimony to a couple of things. I think, firstly, we've got the fact that Office 365 is spread out across the world in 17 different data center regions. And it's got tons and tons of machine resources. A lot of those machine resources were actually put in to deal with artificial intelligence and machine learning. But these resources are available to be repurposed. And one of the interesting aspects of Office 365 is that the degree of automation that exists within the service allows Microsoft to swap workloads to machines very, very, very quickly. So what we're seeing at the moment is two things, really. Firstly, Microsoft is rebalancing workloads. It's rebalancing the resources that are dedicated to the demand created by customers. And the second thing, internally, within applications, it's also turning things back. So the applications themselves are not generating as much demand as they did before. And the logic here, of course, is that it's much more important to provide a service to end users who want to do something rather than a background process that needs to do something because it's nice to do it. You can argue about whether or not that stuff-- that background processing is really important and needs to have equal priority, but the fact is that an awful lot of the changes that Microsoft have made and continue to make are [INAUDIBLE]. So for example, the cache to data is not refreshed as often. In Teams, you have to type more characters before an address is validated. Background processing for things like SharePoint, for the data loss prevention is now slower than it used to be, because when you upload new documents to SharePoint, those documents need to be processed. They need to be indexed. That's done by a background crawler. Part of that crawling process is to check the document against data loss prevention policies. That used to happen quite quickly, maybe a [INAUDIBLE] guarantee it'd maybe happen within 10 minutes-- a new document being uploaded. Now it might take half an hour or an hour. You also see things like Stream has reduced its video resolution down from 1080p to 720p from being recorded, so I don't think that's a great problem. OneNote has been made read only, and I think that's a huge problem. But the point is that, collectively, there's a huge number of changes to be made across Office 365 with the intention that users don't notice that things are really all that much different. But collectively, all the changes that are made to rebalance hardware, to reassign resources, to tune back background processing, and so forth, and so on is-- it'll release resources to take on the extra workload. Just to give you a feeling of how the amount of background processing that goes on inside of Office 365, if you take Exchange Online, there are well over 200,000 mailbox servers running Exchange Online, and maybe 250,000 today. Each of those servers has maybe 100 background systems running. And they're running to do things like apply retention policies, to count the number of users that are in Office 365 groups and distribution, this, and so forth and so on. [INAUDIBLE] piloting-- [INAUDIBLE] check calendar meetings. You name it, there's a background assistant [INAUDIBLE]. Those systems are all being tuned back, and so if you think of-- if you tune back the demand on an assistant by 50% or 100 assistants, that's a lot of CPU cycles just released. If you scale it up to 250,000 servers, that's a hell of a lot of cycles that you've released. So we're not going-- Microsoft is obviously not only making maximum use of its existing resources. It's also bringing on new servers, new storage. This is an ongoing process. Microsoft is a very, very active group of people, who try and predict demand. They couldn't have predicted demand caused by the pandemic, but there are lots of people who are trying to predict what's likely to be the growth curve for Office 365 over time as they open up new data central regions in individual countries to have the data sovereignty, and just to handle the ongoing load created by new users. Every month for the last three-odd months, Office 365 has taken on about 3 million new users. So this is an ongoing process. It's something that never really stopped, and Microsoft is out there buying tons of servers, tons of-- tons of storage and commissioning them as quickly as possible. But as they are moving faster than ever before to try and keep up with demand, clearly, we are going to see over the next while, we're going to see a reduction in background processes that we've just described. You're also going to see probably some more transient errors occurring. You may not them as a user, because things like, for example, Cache Exchange Mode in Outlook will hide this from you for email. But if you're a programmer, or if you were something like an ISV like Crest, and you're running migrations, you're likely to see a lot more transient errors pop up as strain goes on, because, of course, when you're doing migrations, a lot more data is moved around than almost any other time. And there's a lot of read activity. There's also a ton of write activity, and that's the kind of thing where you'd see those type of errors. And we've seen that with being reported by quite a lot of ISPs. I think, also, we'll see Microsoft [INAUDIBLE] back a little of their focus on machine learning, because quite honestly, right now, instead of trying to get artificial intelligence and a ton of machine learning into as many places inside Office 365 as we can fit, I think it's a matter of keeping the lights on, keeping everything going, and making sure that they deliver the service to users as committed. In terms of numbers, I think the last time we saw a public number for Microsoft in October is that 200 million daily active users for Office 365. With the predicted growth, we would expect to see another figure given to us in their next [INAUDIBLE] results, which would be around the 28th of April, I think. [INAUDIBLE] maybe 25th. We'd expect to see six months growth of about 3 and 1/2 million to 20 million. I would take this up to 220. With the extra growth, it could be at 240 million daily active users. That's the scale of things that Microsoft is trying to deal with right now. It's kind of interesting, to say the least. OK, so that's what I have to say about what's happening at the moment. If anybody's got any questions, I'm willing to take them. Otherwise, I'll plunge in and talk about where Microsoft is going with Office 365. Hey, Tony, I don't see any questions specifically right now for this, for what you just talked about. But for the attendees, I did-- in the chat, I did-- if you scroll up, you'll see I added two links to two of Tony's articles. One is about Microsoft imposes restrictions on Office 365 services to ease demand, and the other one is Office 365, [INAUDIBLE] temporary scaling problems. So some of the things Tony just spoke about that you can access his articles there as well and share them around with your colleagues. Thanks, Jennifer. OK, so let's talk about Microsoft 365 and the substrate and stuff like that. Actually, Tony, can I interrupt one second? Yeah. One question did come in, and so in order to make sure we ask-- answer these, let me ask this one. Regarding stress to the online services, are the assistants you mentioned that support the 200,000 or so Exchange servers running Exchange as well, or are they VMs or containers providing compute resources? No, that's a great question. So Microsoft does not use virtual machines inside Office 365 at all. It doesn't use containers. It uses good old fashioned physical hardware, and there are two major reasons for this. The first is that their provisioning model is very simple. They take servers offline. They reduce the servers down to bare metal, and then they pour and build whatever workload package they want to support on it, like an Exchange server or a SharePoint Server or Teams, whatever. They just pour it on, and that server goes on-- that package goes on with all of the updates, the latest bill, whatever it is. So they've got a very, very simple model. There's not a case of going round to 250,000 mailbox servers and applying the latest cumulative update. That is not what happens. You have about 10% of the server population going through that kind of upgrade cycle on an ongoing basis. It's probably a little bit less now, because they're probably-- they're keeping more servers online. This allows them, Microsoft, to say that they have all of the servers are 99.9 something high figure percentage of the servers all running the same software. It's very important to get consistency across Office 365. But they will always-- they'll tell you that there's a few renegade servers out there that might have slipped through the cracks. But the second reason why they don't use virtualization, even though they have HyperV, they have their own product here, is because simplicity. Simplicity is absolutely critical when you're running operations at this kind of scale. And if you don't talk to the Office 365 architects, the guys who actually have to run this monster infrastructure, they'll tell you that there is no way they want to introduce an added layer of complexity, which virtualization represents into their operations. They'd like to keep things really, really simple. So that's the reason why they don't run virtualized servers, and they don't use containers. Now things might change in the future. That's always possible. I think Microsoft will keep an eye on the economics of the situation to make a decision about where they make future decisions. I have heard some people speculate that containers might be a good thing to do. But again, they have to weigh all of the other complexities of bringing in additional layers, introduce this into their operating model, and figure out whether or not they can get more advantage by introducing that added complexity. Can they do it by keeping it simple, as they do today? That sound reasonable? Yep? Any other questions? That's it for right now. Thanks, Tony. OK, everybody's gone to sleep. Great. OK, so the next section I wanted to talk about where Microsoft is going with this cloud off the system, what's really-- what's really driving them. I took a time to look back, because we're approaching the second decade of 365. Those of you that might know a little bit about the history will know that it was launched in 2011, June, 2011, in New York City, as it happens, by Steve Ballmer. And so there's a lot of experience being gained and a lot of scale as we've obviously discussed. So where are they going? Well, so if we will look back, and we looked at where Office365 came from, it had its roots very much in the on-premises world, Windows Server world. You guys who run Windows Server's Active Directory, you know what it's like. And basically, an on-premises deployment is connected by an internal network. But the actual links between the application themselves are pretty disconnected. There's no-- it's impossible, really, to build much dependency between, say, Exchange 2016 and SharePoint 2016. That just doesn't tend to happen in the on-premises world. It's just too difficult to make sure that everything has got the same service pack, cumulative update. They've got the-- you've got the various authentication, stuff like that between the various applications to make sure the data can parse together. So the on-premises world, when we look at it, it's very, very different than what we see in the cloud. It was the same in the early days. 2011, we had Exchange Online, which was basically Exchange 2010 with a little bit of added administration. Same for SharePoint 2010. Same for Office Communication Server, and then Lync Server. You know, these were classic on-prem applications, engineered form prem, developed, used, deployed, managed on prem, and developed and deployed in a way that there wasn't a lot of interaction. But what's changed over the last nine or so years is that you're seeing that Office 365 is really a software tool box. And what this means is that Microsoft's focus now is leveraging all of the bits of Office 365 as much as possible and in as intelligent a way as possible to build new applications and new experiences to deliver to customers. Though it's a different world, and it's a different kind of environment for people to work. And that's a different kind of environment for engineers, particularly, to work in, because you might have people who work for 10 or 20 years on Exchange or SharePoint or whatever. And their view of life is very much OK. It's whatever Exchange or SharePoint could do. And now it's a case of, well, Exchange isn't really the center of the ecosystem anymore. Exchange is much more a bit player within Office 365. It's a very important bit player from a storage perspective, which we'll talk about in a little while, but also from the delivery of email services. The same thing for SharePoint-- SharePoint, if you go to a SharePoint show or whatever, you'll see tons of people that have built a business in the SharePoint ecosystem with add-ons for SharePoint, et cetera, et cetera. That works really well, the SharePoint on-prem, but in the cloud, SharePoint is really just the Office 365 document management service. And that kind of change takes a lot of getting over with, but that's what Microsoft is trying to do. They have this infrastructure, and now they are focusing on how to develop it further and, most of all, how to exploit what they have gathered. So this diagram I did after listening to a talk by Jeffrey Stover at last year's Ignite. And Jeffrey is a good mate. He's-- he and I used to work together in Digital Equipment. He is the architect or one of the big architects for Office 365 with a specific focus on artificial intelligence, the introduction of artificial intelligence, machine learning, and those type of techniques into Office 365. Jeffrey is probably best known for-- he's the father of PowerShell. He's the guy who brought that scripting and automation capability to Windows Server. But anyway, Jeffrey was trying to explain to people what the substrate was. Now the substrate is a term that Microsoft has been using for a number of years but really didn't use outside Microsoft when they're talking to customers. But the way Jeffrey was trying to describe it is he said-- you could think of Office 365 an operation system, where you've got basic object, basic servers, and basic apps. And he said, this is what we rebuilt with Office 365, so that we have piles of different objects that people interact with-- just documents and email and meetings, tasks, whatever. Those objects are exposed through various services, some of which are very well known, like the Applications Exchange in SharePoint. Some of them might be apps like Microsoft Graph API, Exchange Web Services, [INAUDIBLE], et cetera. And then you've got this thing in the middle, which is this is this graph that's built of-- graph of trillions of data points in it, which marks all the connections and interactions between all of the different services. And then on top of it, you've got all these apps, which is what end users access to actually interact with all the information behind. The basic thing is to build the substrate, build the objects, build where they're stored, where they're indexed, where they're processed, where they're dealt with in common. Now there's a technical challenge here, of course, straight up, because when you think of it, well, OK, so Exchange's [INAUDIBLE] database, EFC database engine. SharePoint uses Azure [INAUDIBLE]. Teams uses Cosmos DB. How are we going to bring all those together? How do I get a, for example, an indexing capability across all of those platforms so that I can perform a search and find things? How do I put stuff in one place so that retention is processed in a consistent manner. And what's really happening over the last while is that the substrate has been developing based on Exchange storage. EFT has always been a very flexible database engine. It's an engine that can store just about anything, as you know. We look at what email people send and the attachments they add to email. And what's happening is that you see that stuff that people are working on are now stored in mailboxes. Some of them are in parts of user mailboxes that were visible. Some of them are parts of user mailboxes that are invisible. And some of them are in special mailboxes that are totally invisible, which Microsoft sometimes calls shards. Other people call them phantom mailboxes. But you can see the documents are stored there, if they're created in SharePoint or OneDrive. Teams charts and channel conversations-- copies of them are stored there. You've got the Exchange messages. You got Tasks, the task created by To Do or Outlook, Teams and so forth and so on. Everything is stored now in the substrate. And the substrate has been the one place Microsoft can apply artificial intelligence through, can apply indexing, search against, can apply information protection to secure information and so forth and so on. So it's actually a compelling vision, but it's a vision that could only come about, because Office 365 is becoming closer and closer as time goes by. So in essence, this is it. This is what the influence of the substrate is. We start to get a rationalization of the types of data used across Office 365. Now the rationalization is that people can start thinking of, OK, a document, a document they used here it's used as an attachment in email, used because it documents an item inside SharePoint and OneDrive. OneNote can also be a document. I mean you start thinking about what that document is, and we start to see things like the fluid foundation, which is presenting a vision of what the next generation of those documents might be. You just think of them as XML containers. Inside that XML container, you've got all mass of different type of information. So then you have others things like tasks. Tasks is to a very good example, because when you think about your Outlook tasks, your To Do items, your Task [INAUDIBLE] project, et cetera, et cetera-- lots of different ways to capture a piece of work that needs to be done and record its progress. Now what's happening now is that we're heading very quickly to a single task object, and that single task object will be exposed in To Do. It will be exposed in Outlook. It will be exposed in Planner, and so forth and so on. And this is what's happening all the time is Microsoft works the way through the various types of objects that are used inside Office 365. We talked about the storage. It's a common repository. Don't think of it as Exchange Mailbox Databases. They are mailboxes. They are databases, but Microsoft just refers to it as substrate storage. They don't-- they just think of it as interacting with the database, which is the way we think about it. The access to the data is through a common interface to the graph, and over the last while, we've seen an explosion in the number of graph APIs that are available. People are getting their heads around the graph. They're figuring out that, hey, the graph is a pretty good way of getting that information in a way that maybe we haven't been able to get at before. And it's also given us a lot of connected information, which is also interesting. Instead of using one technique to interact with Teams and another to interact with SharePoint, I now use the same technique, same programming capabilities to interact with both. Then also, because we're building this common repository that allows huge scope to create new applications, you know what the repository is. You know what the data is inside the application, so you know what the API for that data is. So therefore, you can start saying, well, what could I do with this? And this is where the fluid foundation comes in, because the fluid foundation allows us to think about, well, OK, I've got different pieces of application data here, and I make an update to them. The updates all flow back to the substrate, so therefore, that means that anybody who's connected with the substrate can see the update in almost real time. And then the last thing is, because the substrate is drawing information from all around Office 365, it gives Microsoft this huge data set for machine learning. And when it comes to AI and machine learning, the bigger the data set you have, the more precise you're going to be, the less mistakes you're going to make, the fewer false positives you're going to have. There's a lot of business that you'll see so many more edge cases, and so forth and so on. So developing the substrate has been a huge task that Microsoft has been engaged on for probably the last four or five years, and we are really seeing it coming into play now. This is going to be the biggest influence over the long term on how people interact and work with and program against and manage Office 365. So it's not just for organizations like yourselves. It's also for the ISV community, because, of course, the ISV community needs to think about how they can take whatever they're offering, whatever functionality added value, things that Microsoft simply doesn't do, how they can move that forward to deal with a world where the substrate is the predominant influence on everything that happens in Office 365. So it's a really interesting time that we're living in. I think over the next five years, you can see the completion of this vision, and eventually, you might even say that at that stage, there will be one kind of substrate and all of that-- perhaps even the individual application databases will be gone. But who cares, because it's all background engineering. That's all. It's all plumbing. And in the cloud, when you buy a cloud service, you don't really care what the plumbing is as long as the cloud delivers. So that's pretty well what I would have to say on that, so in conclusion, the two sides of the house, we've had an explosion of interest, stress, strain, demand, and Office 365 is a huge challenge. And it seems to pass the test. There are stress points, absolutely. There were little cracks appeared, absolutely. Some of those cracks were painful for people. You bet your bottom dollar. But the point is things didn't collapse. We kept working, so that was pretty good. And going forward, I've said the new applications features you can see over the next number of years, they're going to depend on a combination of machine learning, AI, graph, substrate. That's the magic trinity in Office 365 going forward. OK, Jennifer, that's all I had to say. Now I can take any questions. Yes, yes. I have a couple here that I want to ask. This one I missed earlier-- my apologies. So regarding changes in the strain putting on Office 365 services, you mentioned that machine learning scalebacks. We have a question. Would that machine learning scaleback hurt Office 365 security and compliance services? So any machine learning user on spam, phishing, filters, quarantine items, et cetera? No. No, I don't-- I really don't believe so. Any indication I've got from Microsoft is that they're-- it's kind of the machine learning that's for some of their newer applications that they're working on. I mean they're working on public previews, for example, of how to apply protection encryption to data at [INAUDIBLE]. So let's say you're an organization that's got 250,000 SharePoint documents. To ask a user to go and apply a sensitivity label to each one of those documents would be a task that would be very boring, very mundane, and would never get done. But if you can build an AI and build models that will go-- will understand what your information is, so it's stored in all the different sites across your tenant. You can then have background processes that will go and search for matching documents, all the variations of a document, like a purchase order or a travel request or a HR review or something like that, and apply the appropriate sensitivity label to them. So that's one thing where they'd be scaling back. You might have heard about Project Cortex, which is a big new SharePoint development. Again, that depends on a lot of machine learning to understand about things like, what does an acronym mean? What terms does a company use inside the general course of day to day business? That kind of stuff has been tuned back, because it's still under development, and it's not a production system. A production system's like advanced threat protection, things that anything to do with security, we're absolutely sure that it has not been scaled back at all. That's certainly reassuring, then. Question regarding I think your second half of your presentation, shows this is from Lucas. Hi, Lucas. Lucas is one of our longtime [INAUDIBLE] members. He says, with the shift in moving everything toward ESE and the long history of exchange being a high value target for malicious actors and the condensing of access methods to graph, what efforts are happening in tandem to better secure homogeneous data? Would you like to define what you mean by homogeneous data? You mean the data belonging to what-- an individual? Can we have some more clarification? This is where a face to face meeting really [INAUDIBLE]. [LAUGHS] Yes, if we had had this in Teams. Let me see if Lucas can unmute himself. OK, he says, just data with a common storage methodology. You know, the thing is it's not going to get any worse than it is today. And it's probably going to get better, because, again, you will have extra intelligence that will be able to fill in the gaps where humans make mistakes. That doesn't mean that you're ever going to take humans out of the equation, because I don't think you will. You always need to have somebody looking after machines. Otherwise, we end up with Skynet. But I do think you'll get better protection over the long term. Of course, we'll have to wait and see, but that's my belief. All right. Brad says, to what extent do you think Microsoft will make the information it collects through the substrate available to third parties for analysis or be able to consume at the Microsoft machine learning offerings? Well, you know, the data is all in Graph. The Graph is publicly available. [INAUDIBLE] there are a huge number of public Graph API endpoints. So those endpoints are available to allow people to gather stuff and use them. So for example, you could use Graph Calls to grab all the Exchange storage information, SharePoint storage information, SharePoint [INAUDIBLE] usage information, stuff it into whatever repository you wanted to, like Splunk or anything else. You go and analyze it to your heart's content. You can do that. Those public APIs are all available. All you need to [INAUDIBLE] is a little bit of code. You can even do it with PowerShell. The question might be that some of the more interesting material might not be accessible, and that could be either because Microsoft has made its decision not to make that accessible, or they're working on the API. I mean there's a lot of Graph APIs in beta as well, so you end up with a B1 version and a beta version of-- sometimes you need to go to the beta version of the endpoint to get the-- get extended information. But I think they're really serious about making as much stuff available out of the Graph as possible. So that means that the question then becomes, firstly, what do I know? What do I think I need to know? Second thing is, where can I get the underlying data out of the Graph? Which one of the endpoints are-- which combination of endpoints, because a lot of the endpoints are pretty specific to certain kinds of data? And then the third thing would be, well, where do I put the data when I've retrieved it from the Graph? And then the fourth thing is, well, once I put the data somewhere safe, what do I do with it? How do I analyze it? But the data is there. It's in the Graph. It's going to be in the Graph. That's why it's a single API. That's the beauty of it. All right. If anybody wants to ask any more questions, please put those in the Q&A right now. I'm going to ask one question here, Tony. Given your interaction with clients you consult with through social media, I guess right now, what's the-- what's the biggest thing you're asked about with regards to Office 365 management in terms of this crisis? Is there-- is there any-- is there anything that people are coming to you and asking you about repeatedly right now during this crisis, as it relates to Office 365? Some questions about license management. You know, a company might have made a big investment in licenses, and they've allocated them. And all of a sudden, they need more licenses. They don't want to go and give Microsoft a whole hunk of change, so it's just a-- there's been a number of questions, though. How do I identify people who aren't using licenses, because I'd really like to reuse those licenses. So that's one thing I've seen. I've seen a lot of questions asking about why things aren't working. And some of the reasons why things aren't working is that there are these transient errors, or this Microsoft just thrusted something back. An example of that is that Microsoft did introduce in Teams this notion of suggested entries in your activity feed, which is kind of like information that we picked out of all of the teams that you subscribe to, that you're a member of, that we think would be interesting for you. Well, those type of notifications have disappeared for the moment, because that feature was turned off. They also have to turn off things like the read receipt feature in Teams. In Europe, that's coming back. So people have been asking, why aren't things working? And in some cases, it's a matter of, well, try it again in a few minutes. In other cases, it's, well, you know, that feature, by the way, has just been turned off for the moment. Microsoft hasn't done a great job, I've got to say, of communicating the exact features that have been turned off or the ones that have been slowed down that made some public commentary about things like the definition of beta recordings going to 720 people. At the level of detail that causes people to worry via helpdesk reports, they haven't really given as much information as I would have liked. All right. Yeah, I would think that last one is very pertinent. The first one you mentioned, license management-- are folks concerned about, like, a-- like in a month from now or a couple weeks from now, getting a really big bill from Microsoft? Well, there's a couple of things. The first thing is that they may find that they have a whole heap of users who need to now use a particular thing. So you might, for example, find that you want to have users using-- I mentioned conditional access policies. That's [INAUDIBLE] early on. OK, so what conditional access policies do I need? Well, now do I need Azure AD premium licenses? Let's have a look at what we thought in the past versus what we need to use. Who needs to get that protection. Another thing, too, is that a lot of the features that you'll hear Microsoft making a big deal of are actually Office 365 E5 features. So people say, well, do we need that? Do we really need that, or is it a must have? Is it something that we can live without? And if we need-- if some people need it, well, how many E5 licenses do I need? Should I buy Microsoft 365v5, or do I just need Office365? So licensing is a horrible, complex area. And, you know, you should-- it's like you're probably-- you talk to your bank manager if you're going to get a loan. Go talk to a licensed professional if you're going to sort out your licensing, because they're real now. I think there are-- I think there's a whole group of Gartner analysts dedicated to just license management, so. Yep. That's definitely a complex area. The stuff that you can learn-- you can spend a whole pile of money that you don't need, but the other thing, the other caution I give is that you can actually incur a liability that you don't realize. And that is when the administrator might enable a feature, and they don't realize that that feature requires a particular license. And then there's a lot of places you see in the Microsoft documentation, where they say, well, by the way, we don't have the capability to turn features on and off, because the license of mobile, but we will probably have it in the future, so you can get yourself into a situation where two things happen. One is you've incurred a liability for a whole pile of licenses, and the second thing is that, all of a sudden, when Microsoft does come up with a code to control access to a feature and the license on a license basis, they enable that. You lose a lot of functionality that, perhaps, the business processes were built around that. That also speaks to the need to understand what features you're using inside Office 365. Understand the licenses you need for those features. And make sure that you have those licenses in place and allocated to the right people. Fun. All right, I don't see any other questions coming in right now. Tony, I appreciate your time on this. I think this is really valuable information. I've had a few private chats, folks already sharing their appreciation, so I'm going to just share my appreciation for your time and educating us about this, especially as you mentioned, Microsoft hasn't made a lot of public comments about the service interruptions or what things have turned off. So this has been really helpful. Well, to be fair, they've been keeping the things going, so I wouldn't be too hypercritical. I swear to god, they've kept things going, you know? Yes, exactly. Well I appreciate time. Those on the phone, I really appreciate your time and attention to this. If I can figure out how to get the recording after here, I will attempt to send that out to those who gave their email when they signed in. And then just for one other piece-- two pieces of information, as I mentioned, Tony here is one of our TEC 2019 and our upcoming TEC 2020 speakers. So make sure you check out theexpertsconference.com. A lot of this type of information is what we share through that conference. It's very much a TEC conference, not a West conference. And so make sure you take a look at that. And then we also are planning to do one more of these on April 15. Focus very specifically on your Active Directory, and we'll be bringing in another expert, another TEC speaker, Sean Metcalf, to talk specifically about Azure Active Directory and some of the things you really need to keep your eye on right now. We're all really busy trying to enable that remote workforce. What are those things within Active Directory that you really have to keep your eye on? All right, well thank you, Tony. Thank you, everybody. And enjoy the rest of your day.