2017 is coming to a close and it is a good time to look back and then look forward. Thank you to our customers, partners, and the broader open source community for your participation, 2017 was a year of many accomplishments for LINBIT. We celebrated over 1.6 million downloads of DRBD, expanded into China, and released 4 new technical guides: HA NFS on RHEL 7, HA iSCSI on RHEL 7, HA & DR for ActiveMQ, and DRBD with Encryption. All while helping keep thousands of servers up and running, launching new partnerships, strengthening existing ones, and starting major new technical projects. Thanks to you, LINBIT is a recognized name in data centers everywhere for reliable and fast high availability (HA), disaster recovery (DR), and software defined storage (SDS).
Looking at the market, 2018 promises exciting advances in the world of HA and DR. The following is what I see as the top issues and topics in 2018:
1) Digital Transformation (DX)
Digital transformation is the buzzword of the times. What it means to data centers is an explosion of apps and data. From mobile apps to new or modernized enterprise apps, and data gathered or streaming from thousands of new sources, 2018 will see a pent-up demand for economical solutions to do two things: keep apps up and running, and keep the corresponding mass of data available.
2) Cloud Extends SDS’s reach
Private clouds are becoming more common for large enterprises and with that, traditional storage models are now looking more like legacy systems. SDS has been the go-to tech for storing massive amounts of data in the cloud. SDS provides a lot more scalability over traditional hardware appliances. But SDS is still aimed at tier 2 applications. In 2018, users will begin looking at SDS to replace traditional storage in Tier 1 and Tier 0 environments — like Databases — where speed is a “must”.
3) ROI Continues to be the Litmus Test for Enterprise HA/DR/SDS
HA and DR can be very expensive, especially for larger enterprises who have traditionally bought from proprietary vendors. We have seen a big push toward lower cost, open solutions and I expect this to become more pronounced in 2018. The fact that open source solutions can now match and even exceed proprietary solutions in availability and performance is becoming better understood and more widely accepted. This is in part fueled by public cloud offerings adopting Enterprise solutions based on open source.
4) Protecting New Applications
HA and DR implementations started with standard mission-critical systems like databases, fileservers, and webservers but will continue to expand into new application areas. Examples are messaging data (e.g. ActiveMQ, IBM MQ, RabbitMQ) and SIEM/ESM products (HP Arcsight, IBM QRadar) where uptime is a must. This is especially important as Internet of Things (IoT) unlocks entirely new categories of data (with HA requirements) that feed analytics and artificial intelligence (AI) and machine learning (ML) systems.
Containers have been a game-changer in 2017, further forcing companies to shift from traditional storage to SDS. Container users are beginning to request both persistence for their containerized applications, and software which can manage extremely fast distributed storage. LINBIT has responded by creating both a Docker Driver and a Flex Volume Plugin for Kubernetes, and recently has become Openshift Primed Certified. Other vendors are doing the same. Working with the container application folks to ensure that SDS extends to containerized environments, while providing the performance that users expect will be a 2018 trend.
System infrastructure has advanced over the past few years, thanks to virtualization, cloud software (OpenStack, Open Nebula, and hybrid cloud integration), hyperconverged systems, fast interconnects like InfiniBand or Omnipath, new storage-class memory starting with Flash and now with NVMe adoption, composable infrastructure, etc. They each impact HA requirements in their own way, enablig new capabilities and presenting new challenges. 2018 will see the beginning of standardization in this area.
7) Ease of Use and Management
By nature, HA and DR are first and foremost about predictable and trustable reliability for your applications. Any of you who have been in the cockpit knows that HA & especially DR can be complex when constrained by specific project requirements. Manageability and ease of use become increasingly important as HA setups scale.
A look to 2019
8) New Algorithms
The increasing size of data and the highly distributed elastic apps means that a) some data is likely to be streaming at any given time and b) data availability must scale to new heights. 2019 will see visible progress with new HA algorithms like “erasure coding”, both at the R&D level and in enterprise environments.
Let me know what you see from your vantage point and how we might join forces to push it all forward.
Happy holidays and a happy new year!