A topic given much attention with a lot of hype surrounding it were the new Multi-geo capabilities in Office 365. With not the most intuitive of names, slightly disappointing to those of us interested in the purer performance aspects of the platform, yet of equal importance relating to more compliance aspects of data management, this proved to be a very hot topic. With Multi-Geo capabilities in OneDrive and SharePoint Online, an organisation can expand its Office 365 presence to multiple geographic regions and/or countries within an existing tenant. It allows for provisioning and storage of data-at-rest in the geo locations that you’ve chosen to meet data residency requirements. That’s pretty much where it stops, so this is all about data compliance and meeting data sovereignty requirements rather than any performance gains to be had utilizing this capability.
A decade of designing global SharePoint farms has taught us that endeavoring to strike the right balance of hosting SharePoint services locally for performance gains, whilst managing a central configuration of information and storage of content is much of a dark art. It is an ambition rarely achieved without deeper consequences. The mindset of “being closer to the user equals being faster for the user” bares little if any importance to us when considering connectivity to cloud services in today’s world.
This is where we can introduce the “Microsoft Global Network” – very much an unsung hero in my opinion. This exists today, and has done for a very long time, but is rapidly expanding and evolving for the better. For organisations utilizing Office 365 or an Azure hosted service, this is all about traffic entering the Microsoft global network through strategically placed Microsoft edge nodes. These edge nodes are directly interconnected to more than 2,500 unique ISPs through thousands of connections in more than 130 locations. Microsoft has invested billions of dollars into this network. The interconnection strategy optimizes the paths that data travels on the global network. Organisations get a better network experience with less latency, jitter, and packet loss with more throughput. Direct interconnections give customers better quality of service compared to transit links, because there are fewer hops, fewer parties, and better networking paths.
The image below illustrates the more traditional approach when considering connectivity to cloud hosted services.
Taking advantage of the Microsoft global network, we no longer need to worry about the interconnections, we just need to focus on the final mile of connectivity. This point however, should not be overlooked. To ensure great performance with any Azure hosted application, this needs to be done right.
Looking at some real-world examples with clients who use our intranet-as-a-service platform, Unily, doing this the right way can mean we’re able to host Unily centrally in a single location such as West coast USA and support clients across Europe, Asia and beyond with instantaneous access to content and rapid response times.
Connecting to the Microsoft global network as soon as possible when leaving the client network is the number one goal.
As mentioned earlier, Microsoft peer with over 2,500 ISPs globally to pass traffic from them onto the Microsoft network and back. This network is optimized to get your traffic to and from its destination as quickly as possible and covers an array of Microsoft traffic, regardless of the specific Microsoft technology. It’s important to understand these peer points which can be found in this great article about Office 365 Connectivity. It also explains how we can ensure your traffic is only on the Internet path for a very short period of time until your ISP hands it to Microsoft’s network.
The important thing is to ensure that the peering is happening in a sensible place such as London for a business in the UK, the peering is happening in a reasonable time and that the ISP are doing the best job possible in routing traffic to the Microsoft global network. An ISP may not peer directly with a Microsoft exchange but as long as this is done quickly then there’s no issue.
A bad example would be if your ISP in the UK was resolving to the Microsoft network and accessing an application, like Unily, hosted in an Azure data centre in the US, routing through the ISP network to the East Coast, USA. It’s important to be able to have your Networking and Infrastructure teams recognize this and have your ISP speak to Microsoft directly to set up peering in the UK. As a Microsoft customer, you can flag this with your account team and they can speak to the ISP on your behalf.
In summary, to take advantage of this network locally to its fullest, applications hosted in Microsoft data centres will leverage the optimizations at the edge of this network. This would apply to SaaS solutions built on the Microsoft intranet platform services, such as Unily, as well as Microsoft’s own SaaS offerings such as SharePoint/OneDrive. Ensuring good access to Microsoft's global network offloads the responsibility of optimising your network traffic for the majority of its journey and allows you as an organization to simply focus on that final mile of network connectivity and internal networking infrastructure.
Talk to us to learn more about how Office 365 can support your digital workplace.