When I returned to Forrester in mid-2010, one of the first blog posts I wrote was about Oracle’s new roadmap for SPARC and Solaris, catalyzed by numerous client inquiries and other interactions in which Oracle’s real level of commitment to future SPARC hardware was the topic of discussion. In most cases I could describe the customer mood as skeptical at best, and panicked and committed to migration off of SPARC and Solaris at worst. Nonetheless, after some time spent with Oracle management, I expressed my improved confidence in the new hardware team that Oracle had assembled and their new roadmap for SPARC processors after the successive debacles of the UltraSPARC-5 and Rock processors under Sun’s stewardship.
Two and a half years later, it is obvious that Oracle has delivered on its commitments regarding SPARC and is continuing its investments in SPARC CPU and system design as well as its Solaris OS technology. The latest evolution of SPARC technology, the SPARC T5 and the soon-to-be-announced M5, continue the evolution and design practices set forth by Oracle’s Rick Hetherington in 2010 — incremental evolution of a common set of SPARC cores, differentiation by variation of core count, threads and cache as opposed to fundamental architecture, and a reliable multi-year performance progression of cores and system scalability.
Geek Stuff – New SPARC Hardware
The SPARC T5 is an evolution of the current SPARC T4, which has already claimed a series of impressive performance records since its introduction in late 2011. The T5, moving from the T4’s 40 nm process to a 28 nm fab process, doubles the core count from 8 to 16, each core with the same 8 threads of the T4, improves I/O and memory bandwidth by moving from PCIe 2 interfaces to PCIe 3, and doubling the number of memory controllers from 2 to 4. Along the way they also increased the maximum clock from 3.0 to 3.6 GHz. Oracle has also increased the number of CPU coherency links so that the T5 can be assembled into glueless 8-socket systems. Although Oracle has not formally announced systems, you don’t need to be a rocket scientist to predict that 2 and 4 socket T5 systems will be announced shortly, and an 8-socket system will follow.
Oracle has been somewhat cagey about the “enterprise” variation of the T-series CPU, the M4, intended to anchor the replacement for the current M-series servers. All indications point to a mid-2013 announcement of a CPU that has fewer of the T-series cores with reduced thread counts but massively scaled cache and more sockets in new servers with memory configurations that could be as large as 32 TB. While these new M-Series systems will be interesting as engines for the largest Oracle environments, the T5 systems will be such a significant performance improvement for Oracle that they will probably subsume much of the demand for previous generation M-Series systems. Consider an 8-socket T5 system with 128 3.6 GHz cores in (my guess here) a 5U – 8U enclosure – this is a lot of compute power, capable of running at least 90% of the existing Oracle UNIX workloads, and with Solaris 11’s strong virtualization capabilities capable of being sliced and diced as needed into a large number of VMs for a very flexible enterprise computing resource. Expect T5 modules to rapidly become part of the Oracle Engineered Systems portfolio.
What Does It Mean For SPARC And Solaris Users?
So, geeky technology details aside, what does this mean for current or prospective Oracle system customers? In a nutshell, SPARC/Solaris customers are looking at a predictable long-term future of improved and very competitive performance and price-performance scaling for Oracle hardware, especially when coupled with Oracle software in its “engineered systems.” Oracle and IBM will remain in a strong competition for performance and price-performance across a variety of workloads, and competitive pressures will ensure a decade of strong systems platforms for high-end UNIX workloads regardless of the eventual fate of HP’s Itanium platforms.