Data Center Operations Transformation – Process
Efficient processes and maximum utilization of the physical and human resources within the data center has never been more important.
InCommand facilitates new levels of operations process optimization across real estate, power, cooling, IT, extending even to the links into cloud and third-party organizations.
It uses technologies such as, IoT and Machine Learning to accelerate and de-risk processes used for smart hands, remote hands and Emergency Operations Procedures (EOP).
This is what we call Data Center Operations Transformation.
InCommand provides a single portal that manages all the interventions that impact the physical equipment in the data center. It removes the old painful questions familiar to all data center operators when faced with initiating an action: Is this an IT issue? Is it a facilities issue? Is it a vendor issue? It replaces these with transparency and insight.
What InCommand delivers on the ground starts with the optimization of the data center processes.
In a typical use case familiar to anyone who works in a data center, a request for smart hands may be raised to spin up a server, reboot a piece of equipment, or set up a cross-connect – in fact to execute any kind of request for new deployment for a rejig on existing gear.
Under InCommand this request gets executed, tested and reported back in what is effectively one step.
One way it does this is by integrating mobile and wearable technologies with IoT sensors connected in real time to the Network Operations Center.
Serverfarm’s InCommand provides real time visibility of racks, equipment and conditions directly back to the NOC. So, for example, where an operative is looking at a server, they can communicate directly to a person on the other end of the phone has the same view. They can then outline issues e.g. a particular cable that needs to be pulled on the server.
This streamlines the process which previously involved asking what action is required and waiting for a response.
Instead, this is now completed in real time. Instructions on which cable to be moved or reset button to be pressed are done immediately. This cuts out so much, including time, effort, risk and waste – it is revolutionary.
Emergency Operations Procedures(EOP)
All data centers have experienced a potential or pending service outage. When this happens, InCommand’s optimized processes help de-risk the situation. If, for example there is an urgent need to pull and swap out a blade server, instructions that would previously have been stored on a server would need to be located, confirmed appropriate, printed out and then executed. In any EOP the procedures must be followed to the letter because of the risk of what’s involved. This is often a race against time where there are minutes to go until service goes down.
Using InCommand, the operative receives the automated alert or urgent smart hands request to a mobile device. The alert includes a link to the relevant EOP and physical location of the issue. Once there, the operative can scan the QR code on the device which is recognized by the InCommand system, confirming that it is the right piece of equipment. The operative then begins the procedure directly from the mobile device, carrying out and confirming each step in sequence: carry out grounding and prechecks; prepare the chassis; pull the faulty blade; insert the new blade; re-initialize; confirm power-up. Emergency procedures are now enacted at greater speed and accuracy, with minimized scope for mistakes and based on real-time data.
True data driven maintenance
For maintenance processes, operatives doing rounds are no longer walking around with a clipboard. Instead, using a headset and tablet computers means they are recording as they go – taking meaningful readings and adding value to every task. So, for example, if they spot something that needs an intervention, it can be acted upon quickly. It is no longer something that has to be recorded and followed up at a later date.
For example, where an operative doing rounds spots an issue, a correct maintenance ticket is raised immediately. This goes directly into an algorithm which may create and alert that says: ‘this is a recurring problem and was spotted previously, on this date.” The system then initiates an instruction to bring forward the preventative maintenance intervention.
This is real time management which is effectively run by a fuzzy logic routine which informs when is the right time to maintain this piece of equipment.
Among the outcomes that transform the operation is the ability to track the amount of time spent on preventive maintenance; the percentage of time spent on tickets; how long each ticket takes to close. The outcomes include additional data and analytics on how long it has taken to work on a particular type of ticket. How long that ticket type remained open on average. How many people had to handle it? How can this be reduced? This data can remove risk and improve operations as Incommand’s machine learning (ML) algorithms constantly learn about redundancy and system behaviour. As the training data set grows monitoring becomes more accurate and interventions more timely as predictive analytics manage alerts and improve meant time to service or repair.
IoT enabled monitoring
Where previously equipment had to be hard wired, Serverfarm is now leading the way in the deployment of IoT-based monitoring solutions in the form of wireless sensors with up to a five-year battery life. By communicating directly with InCommand, the sensor brings flexibility compliance and environmental monitoring processes thanks to many more data points generated in real time. From inside a Serverfarm NOC operatives begin monitoring changes in almost real time. If a new rack is deployed at 5pm, by 5.10pm Serverfarm can have a wireless sensor beside it and communicating with both the Serverfarm NOC and the customer NOC if required. Monitoring which once took weeks to set up now takes minutes.
InCommand is providing customers with real data driven decision making and delivering true Data Center Operations Transformation.